Introduction

Melioidosis is a disease caused by Burkholderia pseudomallei, a gram-negative bacterium found in the soil and surface water of tropical and subtropical regions [1]. Statistical modelling of clinical cases and environmental studies estimated that in 2015 there may have been 165000 cases of melioidosis globally and 89000 deaths [2]. With greater environmental disruption from urban expansion, a growing burden of health conditions that increase the risk of melioidosis and, potentially, the impact of climate change, the global burden of melioidosis is anticipated to rise [1,3–5]. Even in well-resourced settings the case fatality rate of melioidosis approaches 10% [6].

Melioidosis usually occurs in individuals with specific comorbidities that include diabetes mellitus, hazardous alcohol consumption, chronic lung disease, chronic kidney disease, and immunosuppression [1]. Patients with cancer are also more likely to develop melioidosis, a result of the immunosuppressive anti-cancer therapy that they frequently receive, although other factors including medical comorbidities and poor nutritional status also contribute to their greater risk [7]. Despite the higher incidence of melioidosis in patients with cancer, the characteristics and clinical course of melioidosis in these patients have rarely been described in detail [1,8,9].

But it is important to define the presentation and optimal clinical management of patients with melioidosis and cancer because cancer incidence is expected to increase globally in the next 20 years, particularly in the transitioning countries which bear a disproportionate burden of melioidosis and where the number of cases of cancer-related melioidosis would be expected to rise [10].

As the case-fatality rate of melioidosis can rise to 50% in these countries, there is also interest in developing strategies that might reduce the incidence of the disease [1,11]. Public health interventions to prevent melioidosis have been unsuccessful and there is no effective vaccine, but chemoprophylaxis has shown utility in some high-risk populations [1]. In one study, no haemodialysis patient in the Northern Territory of Australia receiving daily prophylactic trimethoprim/sulfamethoxazole (TMP-SMX) 160/800mg in the local wet season developed melioidosis, compared to a rate of 17.4% in patients not receiving this therapy; side effects of the TMP-SMX in this cohort were uncommon and mild [12]. The success of this strategy and clinical experience in the Northern Territory has led to Australian national guidelines also recommending TMP-SMX to prevent melioidosis in patients with significant immunosuppression [13]. Daily TMP-SMX (at a dose of 160/800mg) is recommended in individuals with significant immunosuppression (defined as prednisone 20mg daily or equivalent or other potent immunosuppressive drug therapy such as chemotherapy) if they have lived in tropical Australia and have positive serology for B. pseudomallei or if they have a history of melioidosis. The guidelines also recommend consideration of daily wet season TMP-SMX prophylaxis in individuals with significant immunosuppression and negative B. pseudomallei serology who live in, or visit, an area where melioidosis is endemic.

As a result, in the Top End of the Northern Territory, where the incidence of melioidosis is as high as 51.2/100,000 population, patients with cancer receiving immunosuppressive anti-cancer therapy routinely receive TMP-SMX prophylaxis [14]. However, differences in demographics, the burden of comorbidities, and the patterns of environmental exposure means that the incidence of melioidosis is lower in other tropical Australian regions [14–16]. The lower incidence of melioidosis and the potential for severe, and even life-threatening, complications from TMP-SMX prophylaxis, means that the risks and benefits of TMP-SMX prophylaxis for melioidosis in different regions need to be balanced carefully [17].

The incidence of melioidosis in Far North Queensland (FNQ), in tropical Australia, has risen fourfold in the last 25 years; in 2022 the incidence reached 20.4/100,000 population [18]. The region also has a significant incidence of new cancer diagnoses which reached 690.9/100,000 in 2021 [19]. However, it has not been local practice to prescribe TMP-SMX prophylaxis for melioidosis to patients with cancer receiving anti-cancer therapy in FNQ. Local clinicians have not identified a high risk of melioidosis in patients receiving anti-cancer therapy and have concerns about the possible side effects of TMP-SMX prophylaxis–particularly myelosuppression–as well as the potential for medication interactions and polypharmacy. However, with the recent rise in the local incidence of melioidosis the benefits of TMP-SMX prophylaxis may now outweigh these putative risks.

This study had two separate aims. The first was to describe the burden, clinical characteristics, comorbidities, and outcomes of melioidosis in patients with cancer to define the clinical phenotype of melioidosis in this population more precisely. The second aim was to identify if TMP-SMX chemoprophylaxis was associated with a lower rate of subsequent melioidosis in patients receiving anti-cancer therapies. This might help determine if there may be a role for chemoprophylaxis against melioidosis in patients receiving anti-cancer therapies in the FNQ region.

Methods

Results

There was a total of 477 cases melioidosis during the study period. The incidence increased from 4.6/100000 in 1998 to 20.4/100000 in 2022 –the first and last complete calendar year in the study period (p<0.001, r_s = 0.80). The incidence in the CHHHS increased from 1.5/100000 in 1998 to 20.3/100000 in 2022 (p<0.001, r_s = 0.65), but it remained stable in the TCHHS (mean incidence (95% confidence interval): 24.7 (20.3–29.1)/100,000) during the entire study period. In 2022, the incidence in the local First Nations Australians population was 38.8/100,000 compared with 16.6/100,000 in the local non-Indigenous population.

In 446/477 (94%) it was possible to determine if there was an active malignancy at the time of diagnosis of melioidosis; an active cancer was confirmed in 47/446 (11%). The proportion of individuals with melioidosis and an active cancer increased during the study period (p = 0.0499, r_s = 0.09); in the first five years of the study period 0/28 had an active cancer, compared to 29/234 (12%) in the last 5 years.

The characteristics of the patients with melioidosis who did–and did not–have cancer are compared in Table 1. In multivariate analysis, patients with melioidosis and cancer were more likely to be older (odds ratio (OR) (95% confidence interval (CI)): 1.05 (1.03–1.08), p<0.0001) and immunosuppressed (OR (95% CI): 11.54 (5.41–24.6), p<0.0001), but less likely to have diabetes mellitus (OR (95% CI): 0.44 (0.20–0.99), p = 0.048).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of patients with–and without an active cancer–in the cohort.

https://doi.org/10.1371/journal.pntd.0012631.t001

Of the 47 individuals with an active cancer, 37 (79%) had a solid organ tumour and 11 (23%) had a haematological malignancy (1 patient had a solid organ tumour and haematological malignancy concurrently). Individuals with a solid organ tumour were more likely to identify as a First Nations Australian and smoke cigarettes at the time of diagnosis of their melioidosis, but these differences failed to reach statistical significance (p = 0.09 and p = 0.08 respectively using Fisher’s exact test) (S2 Table).

Patients with a solid organ tumour

In 30/37 (81%) with a solid organ tumour, the cancer was already diagnosed and in 7 it was diagnosed during the evaluation and management of the melioidosis. There were 33/35 (94%) who presented with acute symptoms, 2/35 (6%) had chronic symptoms; in 2/37 (5%) the duration of symptoms could not be determined. No cases were thought to have activated from latency.

Most (20/37, 54%) patients with solid organ tumours had metastatic disease. The most common diagnoses were prostate cancer (12/37, 32%) and lung cancer (9/37, 24%). Myelosuppressive anti-cancer therapy had been prescribed to 11/37 (31%) in the 12 months prior to their diagnosis of melioidosis. There were 3/37 (8%) who had received radiotherapy in the 12 months prior their diagnosis of melioidosis, 2 of whom had not received chemotherapy, but both these patients had other risk factors for melioidosis. Indeed, there were only 8/37 (21%) with no other risk factors for melioidosis who had not received myelosuppressive anti-cancer therapy in the prior 12 months (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics and outcomes for patients with melioidosis and a solid organ cancer.

https://doi.org/10.1371/journal.pntd.0012631.t002

Twelve months after the diagnosis of melioidosis, 19/37 (51%) were still alive; 7/18 (39%) deaths were due to melioidosis and 11/18 (61%) were due to the underlying cancer. The 29 patients who completed their intensive therapy for melioidosis received a median (interquartile range (IQR)) of 4 (2–4) weeks of intravenous therapy. Most (22/29, 76%) subsequently received 3 months of eradication therapy with TMP-SMX; four patients had a longer course, 1 patient had no eradication (due to a decision to transition to palliative care) and in two the duration of eradication therapy was uncertain. Skin rash necessitated a switch from TMP-SMX to doxycycline in 4/29 (14%); 1 of these 4 patients represented with melioidosis 4 years later, although it was not possible to perform multi-locus sequence typing to determine if this was relapse or reinfection.

Haematology cohort

There were 11 individuals with a haematological malignancy; the malignancy was diagnosed already in 10 and it was diagnosed in the other during the evaluation of the melioidosis. Patients presented acutely in 8 cases while 3 patients presented with chronic symptoms; no cases were thought to have activated from latency. The most common diagnosis was a lymphoproliferative disorder (5/11, 45%). Most (9/11, 82%) had non-localised disease and 5/11 (45%) were receiving anti-cancer therapy at the time melioidosis was diagnosed. This was myelosuppressive chemotherapy in 3/11 (27%), high dose corticosteroids in 1/11 (9%) and a Janus kinase inhibitor in 1/11 (9%). There were 2/11 (18%) with no other risk factors for melioidosis who had not received myelosuppressive anti-cancer therapy in the prior 12 months (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Characteristics and outcomes for patients with melioidosis and a haematological malignancy.

https://doi.org/10.1371/journal.pntd.0012631.t003

The patients received a median (IQR) of 3 (4–6) weeks of intensive intravenous therapy and either 3 or 6 months of eradication therapy. Second-line therapy was used in 3 cases due to TMP-SMX related side effects. At 12 months 7/11 (64%) were still alive, 2 deaths were attributable to melioidosis.

Association between TMP-SMX prophylaxis during anti-cancer therapy and the subsequent diagnosis of melioidosis

Between 1 January 2008 and June 1, 2023, there were 4400 patients who received myelosuppressive anti-cancer therapy and 368 individuals who received non-immunosuppressive anti-cancer therapy. The incidence of melioidosis in patients receiving this anti-cancer therapy, stratified by myelosuppressive intensity of the regimen, is presented in Table 4. There was no statistical difference in the proportion of individuals who subsequently developed melioidosis among those who did–and did not–receive TMP-SMX prophylaxis for P. jirovecii pneumonia during their anti-cancer therapy. The characteristics of the patients who did–and did not–have immune suppressing anti-cancer therapy prior to the diagnosis of melioidosis (stratified by whether they did–or did not–receive TMP/SMX prophylaxis) is presented in S3 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Proportion of patients receiving TMP-SMX prophylaxis for P. jirovecii pneumonia as part of their anti-cancer therapy and the subsequent incidence of melioidosis (stratified by myelosuppressive intensity of the regimen).

https://doi.org/10.1371/journal.pntd.0012631.t004

Discussion

Despite a significant and rising local incidence of melioidosis, cancer was not a common predisposing factor for the infection in this region of tropical Australia. Indeed, cancer was the sole risk factor for melioidosis in only 2.2% of patients with culture-confirmed melioidosis during the 25-year period, a figure that rose to 4.5% if patients receiving immunosuppressing anti-cancer therapy in the prior 12 months were also considered. Although FNQ clinicians did not routinely prescribe TMP-SMX prophylaxis to local patients receiving anti-cancer therapies, melioidosis was a rare complication in this population. Indeed, there was no statistical difference in the subsequent incidence of melioidosis between patients who did–and did not–receive protocol based TMP-SMX prophylaxis for P. jirovecii pneumonia during the study period. While side effects from TMP-SMX prophylaxis are rare, they may occasionally be life-threatening [24]. Our data therefore do not presently support the prescription of TMP-SMX prophylaxis for melioidosis in patients receiving anti-cancer therapies in this region of Australia.

The proportion of cases in which cancer contributed to the development of melioidosis increased slightly in our cohort during the study period and can be explained by recent changes in the local incidence of the disease. Recent urban expansion is hypothesised to have contributed to a 10-fold increase in the incidence of melioidosis in the city of Cairns, a population where the people are older–and, accordingly, at greater risk of cancer–than the population living in remote FNQ [15,19,25]. Although patients with melioidosis and cancer were more likely to be immunosuppressed, this was often a result of the anti-cancer therapy that they were receiving for their malignancy. The fact that diabetes mellitus was less common among melioidosis patients with an active cancer is likely to be explained by the significant burden of diabetes in the younger First Nations population in the region who are more likely to have diabetes and to live remotely [26,27]. A lower rate of skin and soft tissue disease among patients with cancer is likely to be explained by the fact that SSTI occurs more commonly in patients without any underlying risk factors [28]. The association with chronic lung disease is likely to be explained by the critical role that cigarette smoking plays in the development of both cancer and chronic lung disease. Although patients with melioidosis and cancer were commonly not currently smoking or drinking alcohol hazardously, this may reflect changes in lifestyle precipitated by their diagnosis and treatment [29].

The cancer in the patients with melioidosis and cancer in this cohort was usually already diagnosed and the proportion of different cancers in the patients with melioidosis was similar to that in the local general population [19]. There was no apparent relationship between the patients’ cancer diagnosis and the clinical phenotype of their melioidosis. Although all the patients with lung cancer had melioidosis involving the lung, melioidosis involved the lungs in almost three-quarters of the entire cohort. Furthermore, most patients with lung cancer also had underlying chronic lung disease, itself a risk factor for pulmonary melioidosis [14].

Over 90% of the patients with an active cancer presented acutely with symptoms of melioidosis. Although it is usually impossible to determine if a patient’s presentation represents reactivation of latent infection, the fact that there was no difference in wet season presentation between patients with and without an underlying malignancy suggests that reactivation of latent infection was not common in our cohort.

Death from melioidosis before hospital discharge was more common in patients with melioidosis and cancer than in those without cancer, although this difference failed to reach statistical significance. Relapse of melioidosis was uncommon in patients completing their eradication treatment–occurring in only one patient who received the second-line eradication therapy doxycycline–suggesting that long term secondary prophylaxis may not be necessary in most patients who are able to complete standard eradication therapy. At 12 months, patients were more likely to have died from their underlying cancer (which was often metastatic) than their melioidosis, emphasising the significant contribution that comorbidity makes to the long term outcomes of people who survive their melioidosis [30].

There was no evidence in our study to support routine TMP-SMX use for primary prophylaxis of melioidosis in cancer patients receiving anti-cancer therapy in this region of Australia. This echoes the findings of a study that examined local rheumatological patients taking immunomodulatory therapy [31]. However, prophylactic TMP-SMX has potential benefits beyond reducing the incidence of melioidosis. It is prescribed primarily for prevention of P. jirovecii pneumonia–the reason for its prescription in this cohort–and may have salutary effects against other local opportunistic infections including Nocardia and Listeria [32,33]. Prophylactic TMP-SMX reduces the incidence of bacterial sepsis in patients with transplants and there is also growing interest in the utility of prophylactic TMP-SMX for other common infections in tropical Australia including group A Streptococcus and Staphylococcus aureus [34–36].

However, these putative benefits need to be balanced against the potential for harm from expanded prescription of TMP-SMX. The reported incidence of adverse reactions to TMP-SMX chemoprophylaxis varies significantly in different populations Only 1.3% of an Australian haemodialysis population ceased TMP-SMX due to side effects [12], although an incidence of adverse reactions to TMP-SMX of 41.5/100-patient years and discontinuation rates of 14.5% and have been reported in the Asian rheumatology population [37,38]. While almost all these adverse reactions are mild and less frequent with lower dosing, life-threatening reactions such as Stevens-Johnson Syndrome and toxic epidermal necrolysis are well recognised [39,40]. Pancytopenia may also be more common in patients receiving myelosuppressive chemotherapy, which can be potentially fatal or can, at the very least, interrupt ongoing anti-cancer therapy [37].

Expanded use of prophylactic TMP-SMX has the potential to increase resistance to an agent which has an important role in tropical Australia [41–43]. S. aureus is the commonest cause of sepsis requiring ICU admission in the FNQ region and up to a third of these isolates are now methicillin-resistant with 10% of blood stream isolates resistant to TMP-SMX [27,44,45]. Widespread use of prophylactic TMP-SMX could increase the proportion of resistant isolates further [46]. Evolving antimicrobial resistance–which is driven largely by over-prescription of antibiotics–is an important threat for the continuing success of cancer therapy [47,48]. Infection plays a primary or associated role in the cause of death in approximately 50% of cancer patients [49], and in one meta-analysis the identified pathogen in over a quarter of the infections complicating anti-cancer therapy, was resistant to the standard prophylactic antibiotics that had been prescribed [50].

Concomitant prescription of antibiotics and immune checkpoint inhibitors is associated with significantly worse treatment-related outcomes [51]. This has important implications in a region where immune checkpoint inhibitors are commonly prescribed to patients with lung cancer and melanoma two of the five most common cancers in the region [19]. Finally, there have been significant recent global shortages of some oral antibiotics including TMP-SMX. Expanded use of TMP-SMX prophylaxis has the potential to threaten the supply of the antibiotic for the treatment of important pathogens in tropical Australia and is an additional concern [52].

The study has several significant limitations. The study’s retrospective nature precluded comprehensive data collection on the melioidosis cases prior to 2016 and anti-cancer therapy regimens prior to 2008. The study period was a long one and approaches to the treatment of both melioidosis and malignancy have evolved over this time [14,53,54]. The patients with a malignancy were a highly heterogeneous population–there is a significant difference between a patient with diffuse large B-cell lymphoma and a previous stem cell transplant receiving multi-agent chemotherapy and a patient with localised prostate cancer who is managed with watchful waiting. Melioidosis serology–which is recommended in Australian guidelines to inform melioidosis risk–was not collected, although the limitations of serology in areas that are endemic for melioidosis are well known [55]. Other factors that increase the risk of melioidosis in the patients receiving anti-cancer therapies (comorbidities, environmental exposures) were not collected, however, our findings support the clinical observation that melioidosis is a rare complication of anti-cancer therapy in FNQ. Indeed, many patients received several lines of chemotherapy, so the risk of melioidosis per course of chemotherapy would be even lower than we have reported. Although our data therefore do not presently support the routine prescription of TMP-SMX prophylaxis for melioidosis in cancer patients receiving anti-cancer therapy in this region of Australia, it may have utility in regions where the incidence of melioidosis is higher–such as the Top End of the Northern Territory–and in patients who are also at very high risk of other opportunistic infections. It may also have utility in the prevention of other life-threatening infections, data that we were unable to collect systematically. As ever, it will be a decision for local clinicians to consider the benefits and risks of such an approach [12,17]. If the local incidence of melioidosis continues to rise, TMP-SMX prophylaxis may become more appealing, although short term reductions in the incidence of infection will always need to be balanced against the risk of driving antimicrobial resistance in this high-risk population [47,48]. It was also notable that there was only 1 case of melioidosis–and no cases of fatal melioidosis–among individuals taking a dose of 160/800mg twice a day on two days in a week (a lower dose than the 160/800 daily recommended in current Australian guidelines) [13]. It may be that this lower dose may also be effective for the prevention of melioidosis in jurisdictions that choose to prescribe TMP-SMX chemoprophylaxis for melioidosis.

Conclusions

We have presented the characteristics and clinical course of 47 cases of culture proven melioidosis in patients with active malignancy in tropical Australia. In Australia’s well-resourced health system less than 15% of these individuals died from melioidosis before discharge and disease relapse was uncommon, however, almost a third of the patients surviving their melioidosis had died from their malignancy within 12 months of their infection’s diagnosis. Local clinicians need to consider the risks and benefits of TMP-SMX prophylaxis for melioidosis in cancer patients receiving anti-cancer therapies, but our data do not presently support its routine prescription in this region of Australia.

Supporting information

Acknowledgments

The authors would like to acknowledge the work of the many health care workers involved in the care of the patients.