Introduction

Melioidosis is an infectious disease caused by the environmental Gram-negative bacterium Burkholderia pseudomallei. First recognised in 1911 (Ref. 1) (Fig. 1), the organism is commonly found in the rhizosphere (the layer of soil directly influenced by root secretions and soil microorganisms)2 and surface groundwater of many tropical and subtropical regions3,4, and can infect humans and a wide range of animals.

Figure 1: Milestones in the history of melioidosis.
figure 1

Melioidosis was first recognised in Rangoon in 1911 by the British doctor Alfred Whitmore and his assistant C. S. Krishnaswami, although the name of the disease was coined by Thomas Stanton and William Fletcher. From the time when the aetiological organism was first identified, it has been renamed many times: Bacterium (or Bacillus) whitmori, Malleomyces pseudomallei, Loefflerella pseudomallei, Pfeifferella whitmori, Pseudomonas pseudomallei and, finally, it was officially named Burkholderia pseudomallei in 1992. CDC, Centers for Disease Control and Prevention.

PowerPoint slide

Naturally acquired infections in humans and animals results from exposure through broken skin, inhalation or ingestion of B. pseudomallei5; certain environmental conditions, such as tropical storms and specific occupations (for example, rice farming), are known to increase the risk of exposure6. B. pseudomallei infection can be acute, chronic or latent, although infection usually results in subclinical disease as the majority of immunocompetent individuals can clear the infection. Only those individuals with B. pseudomallei infection who develop clinical symptoms (either acute or chronic) are considered to have melioidosis.

Most cases of melioidosis (85%) result in acute infections from recent bacterial exposure7. The majority of patients with acute melioidosis present with sepsis (a life-threatening, dysregulated, systemic inflammatory and immune response that can cause organ dysfunction) with or without pneumonia, or localized abscesses, regardless of the route of infection. However, the presence of nonspecific signs and symptoms can often hinder the diagnosis and management of melioidosis, which has been nicknamed ‘the great mimicker’ (Ref. 8). Chronic melioidosis is defined as a symptomatic infection that lasts >2 months, and it occurred in 11% of individuals infected with B. pseudomallei in a 20-year prospective Australian study7. The host's immune response to acute infection is both humoral (involving cytokine release, especially interferon-γ (IFNγ)) and cell-mediated, and can completely eradicate or control the infection in most immunocompetent individuals. An unknown percentage of people exposed to B. pseudomallei can develop a latent infection (that is, the infection is asymptomatic and the pathogen is not cleared); activation from latency has been estimated to account for <5% of all melioidosis cases7, but may result in infection becoming apparent many years after exposure.

The case fatality rate of melioidosis is 10–50%6. Of the individuals who survive acute melioidosis, 5–28% experience recurrent infection, which could be due to recrudescence (that is, recurrence) of the original strain, which was not completely cleared and persisted in a dormant state, or reinfection with a different strain following re-exposure6,911. Approximately 80% of patients have known risk factors, mainly diabetes mellitus12 (Box 1). The host–pathogen interplay is complicated by the tropism of the pathogen for a wide variety of cells and its ability to subvert and avoid the host innate immune response13.

Burkholderia mallei is a host-adapted (mainly causing infections in animals) species that originally derived from B. pseudomallei following substantial genome reduction (also known as genome degradation). B. mallei is extremely infectious, mainly to solipeds (mammals that have a single hoof on each foot; for example, horses) but can occasionally infect humans. B. mallei is the aetiological agent of glanders, a disease with similar manifestations to melioidosis. The US Centers for Disease Control and Prevention (CDC) have classified B. pseudomallei and B. mallei (which was used as a biological weapon in World War I)6 as tier 1 select agents because of their biothreat potential (tier 1 select agents present “the greatest risk of deliberate misuse with the most significant potential for mass casualties or devastating effect to the economy, critical infrastructure; or public confidence”)14. No vaccine for either is currently available15,16, which further exacerbates concerns of a possible emerging public health threat.

This Primer summarizes the state of the field in melioidosis research, focusing on epidemiology, pathophysiology (including host–pathogen interactions), diagnostics, screening, prevention and clinical management. In the Outlook, we explore future directions of research in the omics and cutting-edge immunology era, argue whether melioidosis should be recognized as a neglected tropical disease and discuss whether a viable vaccine is on the horizon.

Epidemiology

B. pseudomallei in the environment

B. pseudomallei is well-known to be present in soil and surface water in southeast Asia and northern Australia; however, case reports of melioidosis and predictive modelling studies suggest that it is probably widely present in many countries across the tropics (Box 2).

A consensus guideline for soil sampling for B. pseudomallei was proposed in 2013 with the goal of elucidating the global distribution of the bacterium17. B. pseudomallei is most abundant in soil at depths of ≥10 cm from the surface17; however, during the rainy season it can move from deeper soil layers to the surface, where it can then multiply17.

B. pseudomallei can survive in extreme conditions, such as in distilled (without nutrients) water (for ≥16 years)18, nutrient-depleted soil19 or desert environments20. Outbreaks of melioidosis from contaminated, unchlorinated water supplies have been reported in the Northern Territory, Australia21, and have been associated with chlorination failure (that is, insufficient addition of chlorine to the water) in Western Australia22. B. pseudomallei is also commonly found in unchlorinated water supplies and drinking water in rural areas in Thailand23. Nosocomial (originating in a hospital) infections have been attributed to B. pseudomallei-contaminated wound irrigation fluid, antiseptics and hand wash detergent24,25.

B. pseudomallei has rarely been detected in air. Aerosolized bacteria were first isolated in 1989 (Ref. 26), and in 2015, B. pseudomallei DNA was detected in filtered air using quantitative PCR27. Whole-genome sequencing linked the bacteria found in an air isolate to the clinical isolate from a patient with mediastinal melioidosis (that is, with infection of the midline anatomical structures or connective tissue of the chest)28. Penetration through the skin, ingestion and inhalation are all important routes of infection with environmental B. pseudomallei5. Reported neonatal cases were probably caused by mother-to-child transmission (vertical transmission or breastfeeding)29, healthcare-associated infection29 or community-acquired infection29. Melioidosis is not contagious and human-to-human transmission has rarely been reported6.

Global burden of melioidosis

A 2016 modelling study estimated that there are 165,000 cases of melioidosis in humans per year worldwide, of which 89,000 (54%) are estimated to be fatal3 (Fig. 2). This study highlights that underdiagnosis and under-reporting of melioidosis are a major issue, especially on the Indian subcontinent, where 44% of cases were predicted to occur (predicted incidence for India, Indonesia and Bangladesh are 52,500, 20,000 and 16,900 cases per year, respectively). However, only 1,300 cases were reported per year worldwide since 2010, which is <1% of the estimated annual incidence3. Melioidosis is prevalent in the Northern Territory, Australia, and northeast Thailand, where the annual incidence is up to 50 cases per 100,000 individuals4,7, and the emergence of melioidosis in areas where it was previously absent, for example, in northeastern Brazil, could be explained in part by the increasing recognition of this disease, owing to increased awareness and improved diagnostics30. Although reports of B. pseudomallei isolation from soil and animals in equatorial Africa are limited, they suggest that melioidosis is widely distributed across this region3133. For example, Nigeria is predicted to have an incidence of 13,400 cases per year, which is comparable to incidences observed in endemic regions such as India, Indonesia and Bangladesh3.

Figure 2: Estimated mortality and reported cases of melioidosis.
figure 2

Only Australia, Brunei and Singapore have national surveillance data for melioidosis that are comparable to the estimates. Between 2010 and 2015, there were >100 culture-confirmed cases of melioidosis at a single hospital in Lao People's Democratic Republic yearly247, a number that supports the estimated 420 cases per year countrywide3. However, 20,000 cases of melioidosis per year are estimated in Indonesia, but only 64 have been reported in the country since 1921 (Ref. 248). A large difference between the numbers of predicted and observed cases is also observed in Bangladesh, Brazil, China, India and Nigeria3. This discrepancy could be due to limitations of the model, underuse of clinical microbiology laboratories206, lack of awareness of melioidosis and poor disease reporting systems. Based on data from Ref. 3. N/A, not applicable.

PowerPoint slide

The predicted mortality from melioidosis is comparable to that of measles (95,600 individuals per year) and higher than that for leptospirosis (50,000 individuals per year) and dengue infection (12,500 individuals per year), which are diseases that are considered of high priority by many international health organizations3. Melioidosis can affect all age groups. In prospective studies in Australia and Thailand, the median age of patients with melioidosis was 50 years, with 5–10% of patients of <15 years of age7,34,35.

Risk factors

The most common risk factor predisposing individuals to melioidosis is diabetes mellitus, which is present in >50% of all patients with melioidosis worldwide35,36 (Box 1). Individuals with diabetes mellitus have a 12-fold higher risk of melioidosis after adjustment for age, sex and other risk factors35,36. Other known risk factors include exposure to soil or water (especially during the rainy season), male sex (probably because of a greater risk of environmental exposure), age of >45 years, excess alcohol consumption and liver disease, chronic lung disease, chronic kidney disease and thalassaemia (which probably causes neutrophil dysfunction due to iron overload)6,37. Prolonged steroid use and immunosuppression can also predispose individuals to infection. Nonetheless, >80% of paediatric patients34,38 and 20% of adult patients have no recognized risk factors35,36. Melioidosis in adults who have no risk factors generally occurs in those who have been exposed to a high bacterial load, for example, by aspiration of surface water39. Zoonotic transmission to humans resulting from contact with livestock is extremely rare; only three possible cases have been reported in Australia6.

Mechanisms/pathophysiology

B. pseudomallei is an opportunistic, facultative intracellular, motile saprophyte (an organism that obtains its energy from decaying organic matter) that possesses a remarkable intrinsic array of virulence factors (Table 1) and broad antimicrobial drug resistance (Box 3). B. pseudomallei is highly adaptable, a property that enables it to generate a variety of clinical manifestations, depending on the infected tissue, and to maintain a survival advantage in infected hosts and the environment40. Numerous studies have increased our insights into the pathogenesis of B. pseudomallei infection13,4144 (Fig. 3).

Table 1 Selected virulence factors of Burkholderia pseudomallei
Figure 3: Schematic model of host–pathogen interactions and pathophysiology of melioidosis.
figure 3

Burkholderia pseudomallei secretes N-acyl-homoserine lactones (AHL), which are signalling molecules involved in the quorum sensing machinery that is used to coordinate attacks against the host environment and biofilm formation. The type III secretion system (T3SS) effector proteins are necessary for invasion and escape from the endocytic vesicle; cell entry is aided by flagella, lipopolysaccharide (LPS), type IV pili and adhesins BoaA and BoaB. B. pseudomallei then quickly escapes the vesicle by lysing the membrane using T3SS, T6SS and T2SS. Metabolic flexibility (resistance to oxidative stress), resistance to antimicrobial cationic peptides and ecotin production enable bacteria to survive within an acidic endocytic environment. Effector protein BopA and translocator protein BipD further block sequestration in endocytic vesicles and prevent microtubule-associated proteins 1A/1B light chain 3B (LC3)-associated autophagy. Once free in the cytoplasm, B. pseudomallei replicates, induces the formation of actin-based membrane protrusions and can move via continuous polymerization of host cell actin at polar ends (a process regulated by autotransporter BimA), thereby facilitating spread to neighbouring cells, cell fusion and multinuclear giant cell (MNGC) formation. T6SS and the type IV secretion system (VgrG-5) are essential to this process. Toll-like receptors (TLRs) located on cell surfaces recognize pathogen-associated molecular patterns (such as LPS and flagella) and mediating nuclear factor-κB (NF-κB)-induced activation of the immune response, releasing pro-inflammatory cytokines IL-1β and IL-18. Intracellular inflammasome receptors such as NLR family CARD domain-containing protein 4 (NLRC4) and NACHT, LRR and PYD domains-containing protein 3 (NLRP3) recognize bacterial virulence factors and damage-associated molecular patterns (DAMPS), triggering caspase-1-mediated pyroptosis and further release of IL-1β and IL-18. IL-18 further ensures protective interferon-γ (IFNγ) production (mainly from natural killer cells). Neutrophils, dendritic cells, B cells and T cells are recruited towards the site of infection, and the complement and coagulation cascades are activated. AhpC, alkyl hydroperoxide reductase; BLF1, Burkholderia lethal factor 1; CIS, cytokine-inducible SH2-containing protein; DpsA, DNA starvation/stationary phase protection protein; EIF4A, eukaryotic initiation factor 4A; ER, endoplasmic reticulum; iNOS, inducible nitric oxide synthase; IRAK3, interleukin 1 receptor-associated kinase 3; KatG, catalase-peroxidase; MyD88, myeloid differentiation primary response protein; NF-κBIα, NF-κB inhibitor-α; NO, nitric oxide; NOD2, nucleotide-binding oligomerization domain-containing protein 2; ROS, reactive oxygen species; RpoS, RNA polymerase σ-factor RpoS; SOCS3, suppressor of cytokine signalling 3; SodC, copper/zinc superoxide dismutase; TNF, tumour necrosis factor; TRAF6, TNF receptor-associated factor 6; TssM, type VI secretion system.

PowerPoint slide

B. pseudomallei infection

B. pseudomallei first enters and replicates in epithelial cells of the mucosal surface or broken skin, depending on the route of entry, and then spreads to various cell types. However, the mechanisms of cell invasion and replication are largely similar and are, therefore, discussed collectively (unless otherwise specified).

Epithelial attachment and cell invasion. B. pseudomallei possesses multiple secretion systems, which are evolutionary apparatuses that enable the transport of proteins across cellular membranes in response to the environment and, therefore, host cell invasion. The secretion systems are classified depending on their structure, function and specificity. The type III secretion system (T3SS) comprises a molecular syringe (a structure made of a filamentous needle to translocate effector proteins into the surrounding milieu or cells) that is deployed on close contact with host cells45,46, T2SS is widely distributed in Gram-negative bacteria47 and T5SS secretes autotransporter proteins, which are usually bound to the outer membrane through some adhesin-like proteins48.

Attachment to human pharyngeal epithelial cells was initially thought to be mediated by capsular polysaccharides49 and type IV pili (hair-like structures on the bacterial surface)50. However, internalization into a cell line of human alveolar basal epithelial cells was increased in acapsular mutants compared with wild-type B. pseudomallei51. Furthermore, the type IV pilin protein PilA (encoded by pilA), a subunit type IV pili needed for adhesion to epithelial cells, could also have a role52.

Flagellar motility favours close contact with protective mucosal linings, but flagella are probably not a major adhesin for mammalian cells42,53. Two T5SS adhesin proteins, BoaA and BoaB (Table 1), can enhance adherence. However, double boaA and boaB knockouts show residual binding, indicating that multiple adhesins are required for cell adhesion54. Guanine nucleotide exchange factor BopE, a T3SS effector, causes rearrangement of the host actin cytoskeleton (membrane ruffling) and facilitates ingress55, and BsaQ (a conserved inner membrane T3SS protein) mutants displayed a 30% reduction in invasion42, suggesting that multiple T3SS effectors mediate cell invasion.

Host factors also play a part in epithelial attachment. Protease-activated receptor 1 (PAR1, which belongs to the subfamily of G protein-coupled receptors) is expressed on several cell types (for example, endothelial cells, platelets and monocytes) and promotes B. pseudomallei cell invasion, growth and dissemination56. However, interestingly, PAR1 inactivation had no effect on B. pseudomallei-associated mortality in mouse models, whereas it delayed time to death when the mice were infected with pneumococci56.

Intracellular survival and replication. B. pseudomallei can invade and propagate in both phagocytic and non-phagocytic cells57,58; the bacteria replicate intracellularly, cause lysis or spread to and infect adjacent cells. This process causes acute symptoms, which can vary depending on the tissue or organ infected.

Following endocytosis, B. pseudomallei can be seen in endocytic vesicles and later within the cytoplasm where it replicates59,60. The vesicles then fuse with lysosomes and acidify rapidly61. The T3SS is crucial for vesicle escape before the bacteria can be degraded, as multiple strains mutant for T3SS proteins showed downstream effects, including reduced formation of the actin tail (a comet-like filamentous tail made using actin molecules from the host that the bacteria use for intracellular motility), intracellular survival, cytotoxicity and intercellular spread46,6264. Survival within the endocytic vesicle is aided by an ecotin (a periplasmic serine protease inhibitor) homologue, which is involved in resisting degradation by lysosomal enzymes65.

B. pseudomallei can multiply within phagocytes (including neutrophils, monocytes and macrophages) without activating a bactericidal response57,58. Despite detection of lysosome fusion within B. pseudomallei-infected macrophages (suggesting that degradation of the pathogens can occur to some extent), proliferation of the surviving bacteria ultimately overwhelms the macrophage66. However, macrophages activated by IFNγ (which mediates the immune response to intracellular pathogens) display improved killing of B. pseudomallei, probably via increased activation of inducible nitric oxide synthase (iNOS)67.

In fact, bacterial killing is predominantly mediated by reactive nitrogen intermediates and reactive oxygen species (ROS)67. Consequently, an important mechanism of B. pseudomallei pathogenesis is to suppress iNOS production by upregulating two negative regulatory cytokines: suppressor of cytokine signalling 3 (SOCS3) and cytokine-inducible SH2-containing protein (CIS)68,69. Superoxide (O2) and H2O2 degrading enzymes have been associated with mediating B. pseudomallei resistance to oxidative stress7073 (Table 1).

Evasion of autophagy and cell lysis. B. pseudomallei may trigger autophagy by a T3SS-dependent process that involves the activation of nucleotide-binding oligomerization domain-containing protein 2 (NOD2, an intracellular pathogen recognition receptor)74, resulting in bacterial killing75. However, the exact role of NOD2 might not be clearcut, as another study shows that NOD2 promotes the upregulation of SOCS3 (Ref. 76). Hence, the mechanisms by which NOD2 leads to containment of B. pseudomallei are probably not mediated by cytokine suppression in murine models. Interestingly, polymorphisms in the NOD2 region are associated with susceptibility to melioidosis74. However, the effectiveness of autophagy evasion is influenced by the expression of the T3SS effector protein BopA. Loss of BopA (as suggested by its increased colocalization with microtubule-associated proteins 1A/1B light chain 3B (LC3; also known as MAP1LC3B), an autophagy marker protein) leads to a substantial delay in efficient endosome escape, contributing towards reduced virulence77. T3SS probably plays a crucial part in the evasion of autophagy, as bacteria with mutant BopA are taken up by autophagic vesicles more efficiently and have decreased intracellular survival78. However, the complete mechanisms of autophagy escape remain to be defined.

B. pseudomallei cytotoxicity for certain cell types is also strain-dependent: some strains cause macrophage apoptosis60, some strains cause pyroptosis (an inflammatory form of caspase-1-dependent cell lysis)63 and others cause neither57. Macrophage lysis could represent an escape mechanism for B. pseudomallei once a threshold bacterial load has been reached44. By contrast, apoptosis and degradation of infected neutrophils by macrophages is delayed in melioidosis, favouring bacterial survival79.

Intercellular and secondary spread. Intercellular spread of B. pseudomallei is facilitated by membranous protrusions (generated by bacteria-induced rearrangements of the cytoskeleton) formed by the host cell that extend into neighbouring cells, through which bacteria travel by actin-mediated motility60,80. The autotransporter BimA interacts with monomeric actin at the tail-end of the bacteria, where polymerization occurs81. Nerve root translocation and migration along infected neurons until B. pseudomallei reaches the central nervous system has been supported by animal studies82 and linked especially to the minority bacterial genotype carrying BimABm (B. mallei-like BimA) that is found mainly in Australia83. Such cell-to-cell spread occurring along nerve roots could explain the melioidosis encephalomyelitis syndrome (inflammation of the brain and spinal cord), with brainstem disease following nasal or throat infection, and myelitis (inflammation of the spinal cord) following infection through the skin on the limbs34.

Intercellular spread results in cell fusion and the formation of multinuclear giant cells (MNGCs)84, a hallmark of melioidosis44. Eventual death of MNGCs results in the formation of plaques (one or more MNGCs lyse, leaving a clear zone surrounded by a ring of fused cells) and subsequent damage to host cells, which may serve as a nidus for further B. pseudomallei replication or latent or persistent infection85.

As well as direct cell-to-cell spread, B. pseudomallei can also spread to the bloodstream, causing sepsis, and infect antigen-presenting cells, which then can transport the bacteria to the lymphatic system and contribute to dissemination of infection to secondary sites. However, the exact mechanism of secondary spreading remains elusive. Bacteria also remain viable in dendritic cells, inducing maturation and trafficking to secondary lymphoid organs86.

Latent or persistent infection. B. pseudomallei can remain latent for extended periods before immunosuppression or other host stress responses reactivate bacterial proliferation and melioidosis develops. Reported latency periods have ranged from 19 years to 29 years8789, indicating that B. pseudomallei can enter a dormant state and evade immune surveillance44. Neither the site (tissue or subcellular level) of latency nor the mechanisms by which B. pseudomallei remains undetected are clear90. By contrast, high antibody titres detected in patients years after an episode of acute melioidosis suggest continuous exposure or covert sequestration (bacteria hiding in cryptic sites with downregulation of products)91. B. pseudomallei has been found within the nucleus, which could potentially act as a persistence site for later recrudescence92. Strain variability or small colony variants could also play a part in determining whether latent or persistent infection is established93.

Some B. pseudomallei persistence (and possibly latency) factors have been characterized, including toxin–antitoxin systems (composed of a toxin (protein) and its cognate antitoxin (a protein or non-coding RNA)), metabolic enzymes and adaptive mutations94. By entering a slower growth rate, toxin–antitoxin systems enable bacteria to survive under stressful environments, whereas small colony variants can shift to an acid-tolerant state to survive in abscesses95,96. Furthermore, the host's immune response and selection pressure of antibiotics can contribute to selecting resistance patterns that can also facilitate the establishment of persistent infection (Box 3). Multiple genotypes have been identified within a single infection episode, which at least partly results from genetic adaptation to the human host, including inactivation of virulence and immunogenic factors and deletion of pathways involved in environmental survival97. Thus, bacterial isolates from patients with persistent or recurrent infection show extensive adaptive regulatory changes that favour bacterial persistence, including genome reduction and increased antibiotic resistance. Data do not yet support a correlation between phages and acquired pathogenicity in B. pseudomallei97,98.

Host immune response

Most patients with melioidosis have at least one predisposing risk factor, suggesting that initiation, progression and outcome of the disease are largely determined by the host's immune status12,90. For example, genetic polymorphisms in TNF (encoding tumour necrosis factor), NOD2, TLR4 (encoding Toll-like receptor (TLR) 4) and TLR5 have all been linked to disease severity in patients with melioidosis74,99101. Hypofunctional TLR5 was associated with decreased organ failure, improved survival and a functional cytokine response, possibly mediated by IL-10 (Ref. 102). Interestingly, individuals with the hyporesponsive TLR5 polymorphism display heightened susceptibility to invasive aspergillosis (diseases caused by infection of fungi of the genus Aspergillus) and Legionnaires’ disease (atypical pneumonia caused by Legionella bacteria)103.

Innate immune response. B. pseudomallei activates the complement pathway, but the bactericidal activity of the complement membrane attack complex is hampered by the external capsular polysaccharides of the bacteria104. Owing to its capsule and lipopolysaccharides (LPSs), B. pseudomallei is also resistant to lysosomal defensins and cationic peptides (which contribute to bacterial killing by disrupting the structure of the cell membrane of the pathogen), enabling survival in human serum and within phagocytes44.

Neutrophil, macrophage and lymphocyte recruitment at the point of infection is triggered by activation of pattern recognition receptors, such as TLRs. Despite the possible detrimental effects of excessive neutrophil recruitment105, activated neutrophils play a pivotal part in early bacterial containment106.

TLRs recognize conserved pathogen-associated molecular patterns (PAMPs) and mediate an inflammatory immune response via various signalling adaptor proteins, including myeloid differentiation primary response protein MyD88 (Fig. 3). MyD88 downregulation in experimental melioidosis increases susceptibility to infection as a result of diminished neutrophil recruitment and activation107. B. pseudomallei triggers the upregulation of multiple TLRs including TLR2, TLR4 (and its co-receptor monocyte differentiation antigen CD14) and TLR5 in host cells, leading to the production of pro-inflammatory and anti-inflammatory cytokines via nuclear factor-κB (NF-κB)108110. TLR signalling can be dampened or dysregulated by structural variants of LPSs111,112. In addition, LPS recognition seems to be model-dependent: it occurs solely through TLR4 in murine models, whereas in humans, TLR2 has an additional role113.

Phagosomal escape exposes B. pseudomallei to intracellular TLR-independent pattern recognition receptors, namely, NOD-like receptors, and activates the formation of the inflammasome, a multimeric protein complex that includes a sensor molecule and caspase 1 (Ref. 114). Once the sensor molecule detects B. pseudomallei PAMPs, caspase 1 is activated and rapid pyroptosis ensues115 (Fig. 3). Additionally, the activation of caspase 1 releases active IL-1β and IL-18, which are both increased in patients with septic melioidosis105,116118. IL-18 contributes to IFNγ induction and, therefore, has a protective effect against B. pseudomallei infection105,116, whereas IL-1β has a potential deleterious role owing to excessive recruitment of neutrophils. This interplay supports intracellular bacterial growth, tissue damage and inhibition of IFNγ production105.

Adaptive immune response. Although B. pseudomallei antibodies (due to either past or asymptomatic infection) are common in individuals from melioidosis-endemic regions, their role in developing functional immunity to melioidosis is ambiguous, as reinfection from different strains is possible and, therefore, can occur even in the presence of high antibody levels91.

A strong, comprehensive, cell-mediated immune response is essential for protection against progression of infection and for bacterial clearance119. CD4+ T cells are paramount for B cell isotype switching and for activation of cytotoxic CD8+ T cells and macrophages120. Consistent with this, human survivors of melioidosis display increased levels of CD4+ and CD8+ T cells, whereas a decrease in the levels of these cells is specifically correlated with greater mortality. Moreover, vaccines that evoke an immune response skewed towards the activation of T helper 1 (TH1) cells (which promote cellular responses) provide protection against melioidosis, with potential to generate sterilizing immunity (which protects from the onset of both disease and infection)121,122.

Gradually, granulomas (containing neutrophils, macrophages, lymphocytes and MNGCs) form at the site of infection. Intracellular ‘globes’ of bacteria are seen within MNGCs in a background of acute necrotizing inflammation123, which could lead to the development of granulomas. Unfortunately, the dynamic of granuloma formation has not been studied as extensively in melioidosis as in tuberculosis, but it has been recognized as a source of bacterial reactivation in persistent or latent infections. Most clinicians would advocate active investigation of the nature of granulomas incidentally found in patients from melioidosis-endemic regions.

Unlike in infections caused by other organisms with pathogenicity mechanisms similar to those of B. pseudomallei (such as, for example, non-typhoidal Salmonella124), HIV infection does not seem to be a risk factor for melioidosis or for a more-severe or fatal outcome125. In individuals with HIV infection, macrophages show a dysregulated cytokine response (for example, of TNF, IL-10 and IL-12) owing to a low CD4+ T cell count but retain the capacity for bacterial internalization and intracellular killing126. In murine models of melioidosis, depletion of T cells and natural killer cells (hence a 95% reduction in IFNγ production) did not hamper bacterial control, suggesting substantial redundancy of defence mechanisms and the involvement of macrophages expressing major histocompatibility complex (MHC) class II127. For primary melioidosis, it has been suggested that bystander T cell activation is not required for host survival and could play a more substantial part during the antigen-induced activation phase than during the cytokine-mediated activation127. Paradoxically, a strong CD4+ and CD8+ T cell response was observed during acute infection in patients with melioidosis, and lower responses were associated with increased mortality121. Although the current weight of evidence favours a role for T cells in late stages of infection, little is known about the role of specific T cell subsets in regulating the speed of progression or course of B. pseudomallei infection15.

The inflammatory response. The initial pro-inflammatory response to B. pseudomallei infection is a protective, bacterial killing mechanism. However, a dysregulated cytokine-mediated immune response could result in excessive inflammation with a potentially fatal outcome90. Elevated levels of pro-inflammatory cytokines (such as IL-6, IL-12, IL-15, IL-18, TNF and IFNγ) have been observed in patients with melioidosis, some of which have been correlated with a fatal outcome (for example, IL-6 and IL-18 are considered mortality predictors)116,128. IFNγ production activates T cells and natural killer cells; in murine models, natural killer cells are detected at the site of infection and produce 60–80% of the secreted IFNγ106,129.

Abrogation of TNF or its receptors (in knockout models) results in susceptibility to melioidosis, with increased neutrophil-based inflammatory influx and associated necrosis106,130,131. It is postulated that, in the absence of TNF, a hyperproduction of cytokines and chemokines occurs, leading to septic shock and mortality130.

Anti-inflammatory cytokines (such as IL-10 and IL-4), TNF receptor type I (TNFR1; also known as TNFRSF1A) and IL-1 receptor antagonist (IL1RA) are also upregulated during septic melioidosis. Significant increases in IL-1RA and TNFR1 expression were observed in non-survivors117,132. Late-onset inflammatory mediators are also correlated with clinical outcome and mortality133,134: in experimental melioidosis, macrophage migration inhibitory factor (MIF) impairs bacterial defence, and neutralizing antibodies against high mobility group protein B1 (HMGB1) could be used as an adjunctive therapy to improve outcome. Coagulation also has a role in melioidosis severity (Box 4).

Diagnosis, screening and prevention

Melioidosis is grossly underdiagnosed worldwide3, mainly owing to a lack of diagnostic microbiological laboratories serving the low-income rural populations who are at greatest risk of infection and a lack of awareness of the disease amongst physicians and laboratory staff. Even in good microbiological laboratories, B. pseudomallei could initially be discarded as a contaminant (a commensal or non-pathogenic environmental species), especially in non-endemic areas135.

Clinical diagnosis

The incubation period of acute infections is on average 9 days136, ranging from 1–21 days, although a more severe form of the disease with shorter incubation can occur after inhalation or aspiration of contaminated fresh water137. In patients with melioidosis, the clinical presentation, severity and outcome are affected by the presence or absence of risk factors, the route of infection and the bacterial load and strain, as well as the presence or absence of specific non-ubiquitous B. pseudomallei virulence genes12,83. The clinical spectrum of disease varies from localized cutaneous manifestations at the bacterial entry site with no systemic manifestations to sepsis and death (Fig. 4). Bacteraemia on admission occurs in 40–60% of all patients diagnosed with melioidosis, and septic shock occurs in 20% of all cases. Pneumonia is the presenting illness in about half of all cases. Dissemination of the bacteria to internal organs is common, especially the spleen, prostate, liver and kidneys.

Figure 4: Clinical manifestations of melioidosis.
figure 4

Examples of possible clinical presentations of melioidosis: an MRI of the brainstem and cervical spinal cord with inflammatory changes consistent with encephalomyelitis (arrow, part 1); a ring-enhancing lesion with surrounding oedema in the MRI image indicating cerebral abscesses (arrow, part 2); a CT image of prostatic abscesses (arrow, part 3); a CT image of a mediastinal mass (arrow, part 4); a child with tense parotitis (arrow, part 5); X-ray image of severe pneumonia (arrow, part 6); photo of a subcutaneous abscess (arrow, part 7); and an MRI image of osteomyelitis of the distal femur with surrounding inflammation (arrow, part 8). Clinical images 1–4, 6–8 courtesy of Bart J. Currie, Menzies School of Health, Australia. Clinical image 5 is reproduced with permission from (Ref. 249), Elsevier.

PowerPoint slide

Limitations of clinical diagnosis. Making a diagnosis on clinical grounds alone is very difficult, although in known endemic areas, a patient with suggestive clinical and epidemiological (the presence of risk factors, such as diabetes mellitus and occupational or seasonal exposure) features should be treated empirically with antibiotics that target B. pseudomallei. As the clinical presentation of melioidosis can be nonspecific, the disease should be considered in anyone with a fever in endemic and potentially endemic countries (listed in Ref. 3), in particular in individuals with abscesses (especially in the liver, spleen, prostate or parotid) or pneumonia. However, not all patients have risk factors. Laboratory tests are required to confirm the diagnosis of melioidosis138.

Microbiological diagnosis

Culture. Culture remains the mainstay of melioidosis diagnosis. B. pseudomallei can grow on most routine laboratory media but might not be recognized and could be dismissed as a contaminant, or could be misidentified as other bacteria (such as Pseudomonas spp. and Bacillus spp.) unless laboratory staff are familiar with its appearance138 (Fig. 5). B. pseudomallei is never part of the normal human flora, and its isolation from any clinical sample should be regarded as diagnostic of melioidosis. As B. pseudomallei is classified as a hazard group 3 pathogen and tier 1 select agent139, doctors should alert the hospital laboratory if melioidosis is suspected in an admitted patient to ensure that appropriate precautions can be followed. All microbiology laboratory staff in melioidosis-endemic areas should receive appropriate training regularly and follow local safety standards.

Figure 5: Identification of Burkholderia pseudomallei colonies on three common types of agar.
figure 5

Typical appearance of Burkholderia pseudomallei and Escherichia coli isolated from non-sterile clinical samples. Suspected clinical specimens and suspected bacterial colonies should be processed in a biological safety cabinet. a | B. pseudomallei forms creamy, non-haemolytic colonies that resemble a coliform after 2 days of incubation; by day 4, the colonies are covered by a slight metallic sheen and become dry and wrinkled. b | B. pseudomallei colonies resemble a colourless, non-lactose fermenting coliform after 2 days of incubation; by day 4, the colonies appear dry and wrinkled. c | After 2 days of incubation, the first visible B. pseudomallei colonies are pinpoint with a clear to pale pink colour; by day 4, they become darker pink to purple, flat, slightly dry and wrinkled with a definite metallic sheen. E. coli fails to grow because it is inhibited by gentamicin in the agar. d | Three-disc diffusion antibiotic sensitivity testing: B. pseudomallei is resistant to colistin (or polymyxin) (black arrow) and gentamicin (arrowhead) (although sensitive isolates exist in some areas) and sensitive to co-amoxiclav (red arrow). Parts a–c courtesy of Direk Limmathurotsakul, Premjit Amornchai and Vanaporn Wuthiekanun, Mahidol-Oxford Tropical Medicine Research Unit, Thailand. Part d courtesy of Vanaporn Wuthiekanun, Mahidol-Oxford Tropical Medicine Research Unit, Thailand.

PowerPoint slide

It is crucial that the appropriate clinical samples are collected and sent for culture to laboratories that are familiar with the disease and its causative organism. Blood cultures are the most important sample, as bacteraemia is common. Culture of throat or rectal swabs140,141 and the centrifuged deposit from urine142 are also recommended in all cases of suspected melioidosis, as these could be the only positive samples in some patients. Other samples that should be cultured include pus from abscesses and sputum in patients with pneumonia. Although blood cultures are positive in 50% of patients overall7 and in up to 75% of patients in some reports143,144, this proportion is lower in children38, potentially reflecting that they usually do not have classic melioidosis risk factors. Because cultures have low sensitivity (60%)145, repeating cultures (especially of blood, sputum, urine and pus samples) should be considered in patients with strong indications of B. pseudomallei infection, as it is not uncommon to find subsequent samples positive despite initial negative results. If patients do not improve after 3–7 days of treatment for melioidosis and all culture results are negative, further investigations and a re-evaluation of the diagnosis and treatment should be considered.

B. pseudomallei grows on most routine laboratory media, but more slowly than many other organisms and, therefore, it might be outgrown in samples from sites that normally have a microbial flora; thus, selective media are preferable. Agar plates should be incubated and inspected daily for up to 4 days in suspected cases.

Standard biochemical tests and kit-based identification methods can be used to confirm the identity of colonies. As misidentification of B. pseudomallei with such methods are not uncommon138, the use of multiple methods is recommended. An antibody–antigen binding approach using a monoclonal antibody-based latex agglutination assay is very useful for screening suspect colonies146. Increasing numbers of clinical laboratories use matrix-assisted laser desorption/ionization time of flight (MALDI-TOF) mass spectrometry for bacterial identification, as this method (combined with appropriate databases) can provide an accurate identification of B. pseudomallei147. Numerous molecular approaches for species identification have been described, including 16S rDNA sequencing148 and specific PCRs149, although these tests are often only available in research and reference laboratories. The antimicrobial susceptibility pattern of B. pseudomallei is very characteristic, and in resource-limited areas, a simple disc diffusion antibiotic sensitivity test (to determine the resistance to gentamicin and colistin (or polymyxin) and the susceptibility to amoxicillin-clavulanic acid (also known as co-amoxiclav)) has been recommended to screen Gram-negative rod-shaped bacteria that produce cytochrome oxidase138,143, although it should be noted that gentamicin-susceptible isolates of B. pseudomallei predominate in some regions144.

Direct detection in clinical samples. Because melioidosis can have severe, if not fatal, consequences, treatment should not be delayed by waiting days for culture results; thus, direct detection of the organism in clinical samples could provide a quick confirmation of the diagnosis. B. pseudomallei can be observed with light microscopy (the organism is often described in textbooks as a Gram-negative rod-shaped bacterium with bipolar staining that resembles a safety pin), but light microscopy lacks sensitivity and specificity150. Immunofluorescent microscopy has a specificity approaching 100%, although the sensitivity is <50% compared with culture151.

Alternative approaches using antigen detection or nucleic acid amplification have also been used. A lateral flow immunoassay that detects the extracellular capsular polysaccharides has been developed152, but it has not yet been extensively evaluated and, although it shows good specificity, it seems to have poor sensitivity, especially for blood specimens153. Numerous PCR assays with high specificity for B. pseudomallei have been developed since the 1990s and have undergone small-scale clinical evaluations. The most promising assay targets the T3SS gene cluster154, although the sensitivity in blood samples depends at least in part on an adequate bacterial concentration155. However, these PCR assays are not routinely used for clinical diagnosis in endemic areas, even in high-income countries such as Singapore and Australia: in addition to sensitivity issues, these tests are not cost-effective in providing the timely confirmation of diagnosis, which clinicians need to make therapeutic decisions.

Serology. The serological diagnosis of melioidosis is difficult. Many different assays have been developed for detecting antibodies against B. pseudomallei, but most of them are based on poorly characterized antigens and have never been internationally standardized or subjected to extensive critical evaluation. The most widely used is an indirect haemagglutination test (a simple serological test used to detect antibodies raised against B. pseudomallei). The background seropositivity rates in the healthy population in some endemic areas are very high, presumably because of repeated exposures to B. pseudomallei or closely related organisms156,157. As a result, many patients presenting with fever are misdiagnosed with melioidosis in endemic countries in southeast Asia on the basis of a positive indirect haemagglutination test. By contrast, some patients with melioidosis never mount a good antibody response, perhaps because their immune system is compromised. The indirect haemagglutination test on admission has a reported sensitivity of only 56% in Australia158 and 73% in Thailand157, although in the Australian study, 68% of the patients whose tests were negative on admission subsequently showed seroconversion158. Thus, the diagnosis of melioidosis should not rely on the indirect haemagglutination test.

New assays based on purified antigens are being developed and have undergone small-scale evaluations, with some evidence of improved sensitivity and specificity159161. A protein microarray that contains 20 recombinant and purified B. pseudomallei proteins provides a standardized, easy-to-perform test for the detection of B. pseudomallei-specific antibody patterns162 and could have the potential to improve the serodiagnosis of melioidosis in clinical settings.

Prevention

In northern Australia, basic public health advice is given every year to the general population, and especially to high-risk groups, such as avoiding direct contact with soil and water at the start of each rainy season163. In Thailand, evidence-based guidelines for the prevention of melioidosis recommend that residents, rice farmers and visitors should wear protective gear (such as boots and gloves) if direct contact with soil or water is necessary, only drink bottled or boiled water and avoid outdoor exposure to heavy rain or dust clouds5. The guidelines also encourage cessation of smoking (particularly in those with underlying conditions such as diabetes mellitus) and discourage the application of herbal remedies or organic substances to wounds5. However, the effectiveness of this advice in reducing the incidence of infection has not been proven. Large-scale water chlorination has been very successful in Australia despite theoretical concerns about B. pseudomallei survival in some conditions164. In low-income and middle-income countries, water should be boiled before consumption. Ultraviolet light treatment is effective for remediation of water contaminated with B. pseudomallei and could be recommended in high-income countries in households where individuals are at increased risk of contracting melioidosis165.

However, public awareness of melioidosis in developing tropical countries is limited, and preventive approaches are not always adopted166. A multifaceted intervention at community and government levels is required for successful prevention and is currently being prospectively evaluated in northeast Thailand167.

If high-risk laboratory exposure to B. pseudomallei occurs, post-exposure prophylaxis (PEP) is recommended; high-risk incidents include the exposure of penetrating injuries, mouth or eyes to B. pseudomallei-contaminated materials and the generation of aerosols outside of a biological safety cabinet168. PEP consists of oral antimicrobial treatment with trimethoprim–sulfamethoxazole or, if the organism is resistant or the patient is intolerant, doxycycline or co-amoxiclav for 21 days168. The potential benefit of PEP must be weighed against the fact that trimethoprim–sulfamethoxazole can have severe adverse effects: for individuals involved in a low-risk incident, the decision to begin PEP should be based on the presence of known risk factors for naturally acquired melioidosis. Individuals with known risk factors should be advised to receive PEP, whereas in the absence of known risk factors monitoring is sufficient143,168.

Management

Early diagnosis and the start of antimicrobial therapy specific to B. pseudomallei are crucial for melioidosis treatment. In locations with resources for rapid diagnosis, early implementation of optimal antibiotic therapy and state-of-the-art intensive care facilities for managing severe sepsis, mortality is 10%12. Nevertheless, such resources are not available or are limited in many endemic regions, and in those circumstances, mortality is ≥40%12. The majority of B. pseudomallei isolates from primary infections have the same characteristic antimicrobial susceptibility profiles. B. pseudomallei is susceptible to β-lactam antibiotics (such as ceftazidime, meropenem, imipenem and co-amoxiclav), although the bactericidal activity of these drugs varies, and is almost always susceptible to doxycycline, chloramphenicol and trimethoprim–sulfamethoxazole169171, although these agents only have bacteriostatic activity. Most isolates are susceptible in vitro to piperacillin, ceftriaxone and cefotaxime, but these agents are less effective clinically172. However, B. pseudomallei is resistant to penicillin, ampicillin, first-generation and second-generation cephalosporins, gentamicin, tobramycin, streptomycin, macrolides and polymyxins (Box 3). Of note, clonal groups of isolates susceptible to gentamicin are common in Sarawak, Malaysia144.

On the back of the global concerns of antimicrobial resistance and the already limited options for treating melioidosis, new antimicrobials have been tested in vitro and in animal models, but none can yet replace ceftazidime and meropenem173. Doripenem has minimum inhibitory concentrations (the lowest concentrations that can prevent visible bacterial growth after overnight incubation) similar to meropenem, but ertapenem, tigecycline and moxifloxacin seem to have limited in vitro activity174.

As melioidosis is not a contagious disease, isolation of patients or special precautions are usually not required within endemic areas. However, as few nosocomial infections have been reported24,175,176, healthcare providers are recommended to follow universal precautions177 and standard infection control practices, including hand hygiene178. Potential contamination of the ward and local environment from patients with superficial lesions or pneumonia has been raised as a concern, but such contamination has never been documented.

Formal guidelines for melioidosis therapy, including recommended dosage and duration of each therapeutic phase, have been published by the CDC after a 2010 expert workshop that updated prior consensus-based guidelines168. Antimicrobial therapy consists of the initial intensive phase and the subsequent eradication phase (Box 5).

Initial intensive therapy

Intravenous ceftazidime or meropenem is the preferred choice179; the duration of initial intensive therapy should last a minimum of 10–14 days, with longer intensive therapy for critically ill patients, including those with extensive pulmonary disease, deep-seated collections or organ abscesses, osteomyelitis (infection of bone), septic arthritis or neurological melioidosis (Box 5). The therapeutic response can be slow; the median time to resolution of fever is up to 9 days180, with longer times in patients with deep-seated abscesses. The addition of trimethoprim–sulfamethoxazole to ceftazidime for the intensive phase is used by some physicians for certain types of infection (Box 5) but conferred no survival benefit in studies in Thailand181,182.

Imipenem and meropenem have the lowest minimum inhibitory concentrations against B. pseudomallei and have faster bacterial killing rates than ceftazidime in vitro183,184. The recommendation of meropenem as the drug of choice for severe melioidosis with septic shock is also supported by observational data from Australia185. However, ceftazidime remains the drug of choice for initial therapy for most patients with melioidosis, and there is no conclusive evidence that meropenem is superior to ceftazidime in patients who are not critically ill. In Australia, meropenem is switched to ceftazidime when those patients with severe disease recover and are well enough to be discharged from the intensive care unit to general wards.

Eradication therapy

After the initial intensive therapy, subsequent eradication therapy with oral antibiotics is recommended to prevent recrudescence of the disease or relapse of the patient. Trimethoprim–sulfamethoxazole is the preferred agent for eradication therapy (Box 5), and co-amoxiclav or doxycycline is the second choice. Reports of primary resistance to trimethoprim–sulfamethoxazole in >10% of B. pseudomallei isolates from Thailand and other southeast Asian countries were demonstrated to be inaccurate171.

Based on empirical experience and therapeutic modelling, dosing recommendations for trimethoprim–sulfamethoxazole186 and co-amoxiclav187 in melioidosis are higher than the standard doses used with these antibiotics. Consequently, owing to the long-term course of trimethoprim–sulfamethoxazole required, adverse effects are reported in up to 40% of patients10. In Thailand, this drug is usually avoided in patients who have glucose-6-phosphate dehydrogenase (G6PD) deficiency, owing to the risk of haemolytic anaemia, although patients are not routinely screened for G6PD activity prior to treatment. Rash, gastrointestinal symptoms, hyperkalaemia (high serum potassium levels that can lead to muscle weakness and arrhythmias) and rising levels of serum creatinine (which could indicate renal dysfunction) could require dose modification or cessation of trimethoprim–sulfamethoxazole therapy, which can be replaced by doxycycline or co-amoxiclav. Desensitization (a strategy to safely induce drug tolerance and limit the possibility of a type I hypersensitivity reaction) should be considered for non-severe skin reactions attributed to trimethoprim–sulfamethoxazole.

In Australia, trimethoprim–sulfamethoxazole is the preferred eradication therapy for children and potentially in pregnant women after the first trimester (owing to the risk of neural tube or other congenital defects), whereas in some locations in Thailand co-amoxiclav has been used for eradication therapy in children and in pregnancy. However, dosing with co-amoxiclav is problematic, and acquired resistance is well documented when co-amoxiclav or doxycycline is used4.

Treatment duration

Lengthening the duration of the initial intensive therapy for patients with more-severe melioidosis has contributed to the decrease in mortality in regions with the necessary healthcare resources. A retrospective analysis of patients who were treated according to the Royal Darwin Hospital melioidosis treatment guidelines, which define the minimum recommended duration of initial intensive therapy based on the clinical presentation188, supports a longer intravenous therapy for critically ill patients. The median duration of the initial intensive therapy for such patients in that analysis was 4 weeks and only 5 patients (1.2%) relapsed. Although non-compliance of patients during eradication therapy is common as patients stop taking the antibiotics early, skip doses or do not take any drug at all after discharge, the relapse rates in the Darwin study are low and are attributed to prolongation of the initial intensive therapy188. Studies are required to assess whether future guidelines can include options based solely on intravenous therapy without a long-term eradication phase.

Studies from Thailand showed that failure of the eradication therapy is associated with poor adherence to therapy, more-severe infections (for example, multifocal disease and bacteraemia) and a duration of eradication therapy of <8 weeks189191. These findings support the recommendations for an eradication therapy of 3–6 months (Box 5). Case series reported that selected patients with localized cutaneous disease could be successfully treated with oral trimethoprim–sulfamethoxazole for 3 months without preceding initial intensive therapy34,192,193. However, such a regimen must be restricted to patients who have no signs of sepsis or organ dysfunction, no underlying risk factors and no dissemination of the infection to distant sites, including regional lymph nodes.

Surgical aspects of therapy

Surgical drainage is usually required for single, large abscesses in the liver and muscles and prostatic abscesses194, but it is not necessary or possible for multiple small abscesses in the spleen, liver and kidneys. Septic arthritis (inflammation of the joints due to the infection) usually requires drainage and washout and might require repeated procedures. Other internal abscesses rarely need to be drained as they frequently resolve with medical therapy. Osteomyelitis can be very extensive when diagnosis and appropriate antibiotic therapy are delayed, and in such cases, aggressive and often repeated surgical debridement of the necrotic bone is usually necessary195. However, early long-bone osteomyelitis without abscess formation and vertebral osteomyelitis without epidural abscess might not require debridement. Mycotic aneurysms (caused by bacterial infiltration of the arterial wall) require urgent surgery, often with insertion of prosthetic vascular grafts. Lifelong suppressive therapy with trimethoprim–sulfamethoxazole might be indicated for patients who have received prosthetic grafts for mycotic aneurysms.

Adjuvant therapy

State-of-the-art intensive care management can substantially decrease mortality in patients with melioidosis and sepsis or septic shock196,197, and appropriate guidelines are available198.

Granulocyte colony-stimulating factor (G-CSF, which stimulates the production of neutrophils in the bone marrow) has been used empirically in patients with septic shock. The rationale is to counteract the functional neutrophil defects that are thought to be crucial in the pathogenesis of severe melioidosis. Early observational data showed a significant improvement in survival with G-CSF, but this result was confounded by concomitant improvements in other clinical parameters199. A subsequent randomized controlled trial in Thailand showed no overall survival benefit with addition of G-CSF200; nevertheless, survival was significantly longer in the G-CSF group in settings with limited intensive care resources, and adjuvant G-CSF is still used for melioidosis-associated septic shock in some hospitals with state-of-the-art intensive care facilities. Preclinical studies have shown that administration of clinically available IL-1β blocking agents can protect mice against overt B. pseudomallei infection and mortality105,201. This finding is in line with reports that patients with melioidosis who have diabetes mellitus and take glibenclamide (also known as glyburide) have a lower mortality and attenuated inflammatory responses compared with patients who do not take glibenclamide202 (Box 1). Given the crucial role of immune function in melioidosis pathogenesis, patients with melioidosis-associated sepsis or septic shock could benefit from newly available immune-modulating therapies, such as IL-7, granulocyte-macrophage colony-stimulating factor and anti-PD1 (programmed cell death protein 1).

Quality of life

Recrudescence and recurrence of melioidosis

Recrudescent melioidosis with a return of clinical illness and culture positivity can occur during the eradication treatment period if the initial intensive therapy is ceased too early, internal abscesses have not been diagnosed or adequately drained or if the oral eradication therapy is not adhered to or is ceased too early. Patients with recrudescence will need hospital admission, intravenous parenteral antimicrobials and re-examination for sites of infection.

Once both initial intensive and eradication therapies have been successfully completed, recurrent melioidosis could be due to either recrudescence of the original infection, as confirmed by bacterial isolate genotyping, or a new infection with a different strain190,203. With improvements in therapy, recrudescence decreased from 10% of cases to <5%; new infections in survivors of melioidosis are now more common than recrudescence188.

Long-term sequelae

For survivors of melioidosis, the main determinants of future health are their underlying risk factors that predisposed them to the initial infection. Many of the patients in the Darwin Prospective Melioidosis Study have subsequently died or had a substantial and ongoing disability as a consequence of comorbidities such as diabetes mellitus, chronic kidney disease or malignancy (B.J.C., unpublished observation). The most severe melioidosis-associated disability is residual neurological deficit subsequent to melioidosis-associated encephalomyelitis. Although rare, neurological complications are particularly problematic in children and can range from severe residual quadriparesis (muscle weakness in all limbs) or severe flaccid paraparesis (partial or complete paralysis of both legs) to persisting isolated foot drop (an abnormal gait in which the forefoot drops owing to weakness)34. Limited range of motion, sinus tract (an infected tract connecting a deep-seated infection to the surface or another organ and often discharging pus) formation and joint deformities are also common in patients with bone and joint infections195,204.

Outlook

Melioidosis as a neglected disease

Neglected tropical diseases are understudied diseases that remain endemic in many developing countries around the world205. Melioidosis is not included in most lists of neglected tropical diseases, including the WHO list205, even though it has high mortality and is potentially preventable and treatable3. Thus, efforts by the international research community are needed to raise awareness of melioidosis within the WHO and regional and local health agencies, as well as in the general public in endemic areas. In 2015, the International Melioidosis Society (IMS) was formed by melioidosis researchers to raise awareness and knowledge of the disease amongst all stakeholders, and in 2016, a Research Collaboration Network (RCN) was formed to bring the disease to the attention of public health officials and policy makers in melioidosis-endemic countries. An interactive map and disease information are available on the IMS–RCNwebsite(http://www.melioidosis.info).

Melioidosis in developing countries

Diagnosis. In developing countries, melioidosis is most common amongst members of the rural population, who often have limited access to even simple diagnostics. Furthermore, even if clinical microbiology laboratories are available, they might be underused owing to costs or lack of trained staff206. Non-culture-based methods to diagnose bacterial infectious diseases are increasingly being encouraged207, but they have not been extensively developed and evaluated for melioidosis. Thus, the priorities for improving melioidosis case ascertainment in tropical endemic areas are the establishment of basic clinical microbiology laboratories and education of both clinicians and laboratory technicians about melioidosis and prevention.

Management. To reduce acute melioidosis mortality, the availability and affordability of ceftazidime or carbapenems need to improve. In Thailand, an upper-middle-income country, a 14-day course of ceftazidime or carbapenem costs US$60 or US$1,080, respectively, and is covered by Thailand's Universal Coverage Scheme208. Nonetheless, in other low-income and middle-income countries, these drugs could be more expensive, have limited availability or be excluded from the country's universal health coverage, and patients frequently cannot afford to pay for the drugs themselves209. Studies on how best to allocate and use resources for melioidosis in low-income and middle-income countries are urgently needed.

The One Health Initiative

In 2016, melioidosis was highlighted as a sapronosis, a disease of animals and humans caused by an environmental organism; however, the disease in animals and geographical distribution of distinct human clinical manifestations (for example, parotitis is mainly observed in children within southeast Asia and encephalomyelitis mainly in northern Australia) are not well understood210. Thus, adopting the One Health Initiative (a strategy involving interdisciplinary collaborations of health professionals at the local, national and global levels in all aspects of healthcare for humans, animals and the environment211) will promote cooperation and strategic planning between physicians, ecologists, environmental scientists and veterinarians. Furthermore, interdisciplinary efforts will help to address the sapronotic spread of the disease and to establish effective melioidosis interventions.

To implement One Health Initiative approaches, potential technology-based solutions, such as wireless and mobile technologies for the delivery of health interventions and education and the interactive tools currently displayed on the aforementioned melioidosis website (http://www.melioidosis.info), should be implemented in resource-limited settings. One Health Initiatives for melioidosis could also lead to a broader engagement of organizations and individuals with experience in prevention, surveillance and clinical case management of neglected tropical diseases, as well as individuals with training in economic development, genomics, veterinary sciences, wildlife management, agriculture, molecular biology and bacteriology, ecology, policy and law211. The implementation of this interdisciplinary initiative would combine field efforts addressing both endemic and emerging melioidosis but requires effective global health governance210.

Advances in vaccine development

A vaccine could be a cost-effective intervention in tropical developing countries, particularly if used in high-risk populations such as individuals with diabetes mellitus, even if it produces only partial immunity212. The Steering Group on Melioidosis Vaccine Development has highlighted the need for standardized animal models and for advancing melioidosis vaccine candidates to preclinical and clinical studies, and the challenges posed by bacterial strains and routes of immunization. Furthermore, it recognized that vaccines designed for the military or other populations involved in biodefense have different requirements from those designed to prevent naturally acquired melioidosis in endemic areas213.

No melioidosis vaccine is currently available for human use; some vaccine candidates have been shown to provide partial protection against melioidosis or glanders in murine models of infection212,214,215, but few have been tested in non-human primates or humans216. Live attenuated vaccines induce a more-comprehensive immune response in animal models and are currently considered the best approach for generating protection against B. pseudomallei infection16,217. However, subunit-based vaccines provide a feasible alternative because of their increased safety and potential for large-scale production. Experimental evidence indicates that the combination of bacterial polysaccharides (LPSs or other capsular polysaccharides) with well-defined protein antigens (glycoconjugates) can generate substantial protection against Burkholderia infections217. This rapid advancement in vaccine design and optimization is very promising, and several vaccine candidates are currently being tested. However, a multivalent vaccine containing numerous immunogenic bacterial components will probably be necessary to achieve complete protection, as, in addition to a strong antibody response, both CD4+ and CD8+ T cells are also important in protection against human melioidosis.

Final thoughts

New global environmental sampling studies and improvements in diagnostic microbiology could enhance the understanding of the geographical distribution and burden of B. pseudomallei, and wide-scale whole-genome sequencing together with clinical details could provide new insights into the phylogeny and virulence of these bacteria. The application of such new techniques to isolates with well-characterized clinical and epidemiological metadata could provide further insights into melioidosis. New omics-based technologies will enable a better understanding of how B. pseudomallei evades immune surveillance and can remain latent for many years. The role of the human and environmental microbiota is only just beginning to emerge218 and could offer creative insights into how to tackle the infection in vivo and reduce exposure in terra.

Affordable, effective alternative drugs and new drug targets are needed to reduce mortality, relapse and the duration of treatment courses. More than 80% of individuals with diabetes mellitus live in low-income and middle-income countries, and the numbers are projected to increase by >55% globally by 2050, with tropical countries facing the brunt of this epidemic219. Case clusters of melioidosis have also been associated with severe weather events220,221, which, based on the current estimates of global climate change, might become increasingly common. This combination could cause a surge of melioidosis, particularly in countries where it has previously been under-reported, such as India. Calculating a representative disability-adjusted life year metric could also provide better comparative burden information and prompt regulating bodies to recognize melioidosis as a neglected tropical disease.

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

How to cite this Primer

Wiersinga, W. J. et al. Melioidosis. Nat. Rev. Dis. Primers 4, 17107 (2018).

Further Information

IMS–RCN website