Lung Involvement in Leptospirosis in Tropical Australia; Associations, Clinical Course and Implications for Management

1. Introduction

Leptospirosis is a potentially life-threatening zoonotic disease caused by pathogenic spirochetes of the genus Leptospira. In 2015, it was estimated that globally there were over one million human cases annually resulting in almost 59,000 deaths [1]. The case-fatality rate of patients with severe leptospirosis who are admitted to the intensive care unit (ICU) can exceed 50% [2]. Despite this, leptospirosis remains an under recognised disease and its pathophysiology and optimal management are still incompletely understood [3,4,5,6,7].

Symptomatic leptospirosis typically presents with fever, rigors, myalgia, nausea, vomiting and headache. In most individuals it is a mild illness, but approximately 10% of cases develop severe disease characterised by rapidly evolving multiorgan dysfunction with manifestations that can include hypotension, acute kidney injury (AKI), pulmonary haemorrhage and acute respiratory distress syndrome (ARDS) [5]. Lung involvement is often a harbinger of a more complicated disease course with death almost 10 times more common in patients with lung involvement in some series [8,9,10,11].

However, there is uncertainty about risk factors for lung involvement and how best to manage the patients who develop this complication. Although antibiotics are recommended in patients with symptomatic leptospirosis, the incremental value of antibiotic therapy above optimal supportive care is uncertain [5,12]. Conservative fluid management strategies are frequently employed in individuals with lung involvement although achieving an optimal fluid balance is challenging in individuals who are also often hypotensive and/or have AKI [13,14,15]. Some observational studies have suggested benefit of corticosteroid therapy in individuals with lung involvement, however evidence of benefit in randomised trials is lacking [16,17]. Plasma exchange, plasmapheresis and extracorporeal membrane oxygenation (ECMO) have been used by clinicians to treat severe leptospirosis, particularly when accompanied by pulmonary haemorrhage, however the benefits of these interventions have only been reported in small case series [5,18,19,20]. The role of inhaled or parenteral tranexamic acid for individuals with pulmonary haemorrhage is also incompletely defined [21].

Far North Queensland (FNQ), in tropical Australia, has the highest incidence of leptospirosis in the country [22,23]. Prompt access to sophisticated multidisciplinary care in the well-resourced Australian health system means that deaths from leptospirosis are very rare, even in patients with multi-organ failure requiring ICU care [14,24]. FNQ clinicians recognise that individuals with leptospirosis that is complicated by lung involvement are at greatest risk of death, but what are the characteristics of these patients and what is their optimal management? This study, performed in a high-volume centre where leptospirosis has a very low case-fatality rate, examined the demographic, clinical, laboratory and imaging findings of individuals with – and without – lung involvement. It was hoped that these data, might inform the management strategies of clinicians, less familiar with leptospirosis, who encounter individuals with this potentially life-threatening disease.

2. Materials and Methods

This retrospective cohort study was conducted at Cairns Hospital, the 770-bed referral hospital for the FNQ region. The hospital serves a population of approximately 290,000 who live across an area of approximately 380,000 km² (Figure S1). FNQ has a hot, tropical climate; most local cases of leptospirosis occur in the December-April wet season when the average monthly rainfall can exceed 400 millimetres [25,26].

This is a substudy of a previously published report which examined all adults and children admitted to Cairns Hospital with a confirmed diagnosis of leptospirosis between January 1, 2015, and June 30, 2024 [24]. This study period was chosen as it coincided with the local introduction of an electronic medical record (EMR). Individuals were eligible for inclusion in the study if they satisfied the Australian definition of laboratory confirmed leptospirosis (Table S1) [27]. Culture of Leptospira from whole-blood, with subsequent serovar identification if isolated, polymerase chain reaction (PCR targeting outer membrane protein LipL32) testing and serology (Microscopic Agglutination Test; MAT) were performed at Australia’s Leptospirosis Reference Laboratory in Brisbane, 1390 km away from Cairns Hospital [28]. If individuals had no chest imaging performed during their hospitalisation they were not included in this study.

The patients’ EMRs were examined to collect demographic data and to identify any comorbidities (Table S2). The patients’ symptoms and clinical signs that were documented at their presentation to Cairns Hospital were also recorded. Laboratory data were collected from the Queensland statewide electronic database AUSLAB; these data included the values at presentation (either at the regional referring hospital or at Cairns Hospital) and the highest or, where relevant, lowest values during the patient’s hospitalisation. Specialist radiologist reporting of any imaging was also collated from the health system’s electronic database.

Individuals were said to have lung involvement if they had any new lung parenchymal changes on this chest imaging. Pulmonary haemorrhage was defined as present if there was documented frank haemoptysis or if frank blood was present on tracheal aspirate. ARDS was classified using the Berlin definition based on the partial pressure of arterial oxygen: fraction of inspired oxygen (PaO₂/FiO₂) ratio where this was available [29]. As many individuals did not have an arterial blood gas collected during their hospitalisation the pulse oximetry oxygen saturation: fraction of inspired oxygen ratio (SpO₂/FiO₂) was also calculated [30]. Oliguria was defined as a documented urine output of less than 0.5 mL/kg/hour.

The duration of symptoms prior to antibiotic therapy was noted. If patients developed pulmonary haemorrhage, required ICU admission, needed vasopressor support, required intubation and mechanical ventilation, needed renal replacement therapy (RRT) or required transfer to a quaternary centre for more specialised care, this was recorded. The administration of corticosteroids was documented, as was the agent that was chosen and the duration of this therapy. The total hospital length of stay and the length of stay in the ICU, where relevant, were recorded. All-cause 90-day mortality was documented.

Statistical Analysis

Data were de-identified and entered into a centralised electronic database (Microsoft Excel) and analysed using statistical software (Stata version 18). As many of the continuous variables had a non-parametric distribution they are presented as the median and interquartile range (IQR). Groups were analysed using the Wilcoxon rank-sum test, the chi-squared test, Fisher’s exact test or logistic regression, where appropriate. If individuals were missing data, they were not included in analyses which evaluated those variables.

Ethical Approval

The study was conducted in accordance with the Declaration of Helsinki and approved by the Far North Queensland Human Research Ethics Committee (HREC/EX/2024/QCH/108994) on the 2 August 2024. As the retrospective data were de-identified and presented in an aggregated manner, the Committee waived the requirement for informed consent.

3. Results

There was a total of 111 individuals with laboratory-confirmed leptospirosis admitted to Cairns Hospital during the study period; two individuals had no chest imaging performed during their hospitalisation, leaving 109 individuals who were included in this analysis. Their median age was 39 (IQR: 24-56) years, and 93/109 (85%) were male. Most (65/109, 60%) presented during the wet season; 83/109 (76%) were transferred to Cairns Hospital from a smaller regional health facility for escalation of care.

Leptospirosis PCR was performed on blood in 98/109 (89%) patients and was positive (lipL32 target detected) in 88/98 (90%); whole-blood culture for leptospirosis was performed in 47/109 (44%) and was positive in 29/47 (62%). Overall, PCR of blood or whole-blood culture was positive in 92/101 (91%) who had either of these tests performed. Acute serology testing by MAT was performed in 102/109 (94%) and was reactive in 44/102 (43%). Convalescent serology was collected and tested by MAT in 50/109 (46%). Convalescent serology was frequently performed to infer the infecting serovar, but it established the diagnosis of leptospirosis in 11/50 (22%). It was possible to infer the serovar in 62 individuals; Leptospira interrogans serovar Zanoni (24/62, 39%) and L. interrogans serovar Australis were the most common (12/62, 19%) (Tables S3 and S4). Other elements of the presentation, diagnosis, care and clinical course of the cohort have been reported previously [24].

3.1. Lung Involvement

Overall, 62/109 (57%) had lung involvement; there was no difference in the demographic characteristics or the comorbidities of patients who had lung involvement and those who did not (Table 1). Notably, lung involvement was neither more common in individuals with chronic lung disease nor in those who were current smokers. Individuals with lung involvement had a longer duration of symptoms prior to the commencement of antibiotic therapy than individuals without lung involvement.

Among the 62 individuals with lung involvement, 27 (44%) had respiratory symptoms at presentation: 23 (37%) had a cough, 12 (19%) had haemoptysis and 11 (18%) complained of dyspnoea; 34/62 (54%) had abnormal chest auscultation at presentation (Table 2). On initial laboratory testing, individuals with lung involvement had greater renal impairment, a higher C-reactive protein and were more likely to have an abnormal troponin than individuals without lung involvement (Tables S5 and S6). Individuals with lung involvement usually had a positive blood PCR or positive blood cultures; this included all 41 individuals with lung involvement who required ICU admission. Individuals with lung involvement were more likely to have infection with L. interrogans serovar Zanoni than individuals without lung involvement (Table 1).

The acute parenchymal changes on chest imaging that defined lung involvement were present on initial imaging in 36/62 (58%) and were present on subsequent imaging in another 26 (Table 3). These changes were often multilobar and opacities were frequently alveolar. The presence of multilobar or alveolar changes were associated with a more complicated disease course (Table 3). Nodular parenchymal changes were often present in the most critically ill patients (Figure 1 and Figure 2).

Panels A-C. CXR imaging of an otherwise healthy 24-year-old female who presented with a 3-day history of fevers, but no respiratory symptoms and a normal chest x-ray (Panel A – day 1). She was commenced on ceftriaxone and azithromycin but remained febrile. A repeat CXR (Panel B – day 2) shows blunting of the left costophrenic angle without any focal parenchymal abnormality, but she developed dyspnoea and an oxygen requirement over the subsequent 48 hours. Panel C (day 5) demonstrates the associated widespread patchy alveolar changes. With supportive care, she avoided an ICU admission, made a full recovery and was discharged after a 7-day hospitalisation.

Panels D-F. CXR imaging of an otherwise healthy 37-year-old male reporting a four-day history of cough, fevers and general malaise. At presentation he had type I respiratory failure and required intubation in the Emergency Department before transfer to ICU. A CXR (Panel D – day 1) performed on arrival to Cairns Hospital revealed widespread bilateral patchy airspace consolidation. On day 2 of his admission, blood stained sputum was reported and a repeat CXR (Panel E) showed significant deterioration with diffuse airspace opacities consistent with pulmonary haemorrhage. Refractory hypoxia necessitated commencement of ECMO which he received for 20 days. Panel F, 26 days after his presentation, shows nearly complete resolution of the lung changes. He was discharged after a total of 8 weeks in hospital

CT chest images (A: axial and B: coronal) of an otherwise healthy 30-year-old male who presented with 4 days of generalised malaise, fevers and haematuria. On presentation he was haemodynamically stable and had no respiratory symptoms; however, within 8 hours he deteriorated requiring vasopressor support and supplemental oxygen. Images shortly after deterioration demonstrate diffuse alveolar opacities, ground glass opacities and centrilobular nodules. The patient was admitted to ICU where he required intubation and mechanical ventilation. He made a full recovery and was discharged after a 20-day hospitalisation.

CT chest images (C: axial and D: coronal) of an otherwise healthy 37-year-old male who presented with a 3-day history of cough, fevers and dyspnoea. He was tachypnoeic at presentation but had a clear chest on auscultation, normal oxygen saturation on room air by pulse oximetry and a normal chest x-ray. Within 3 hours of presentation the patient developed worsening respiratory distress and an oxygen requirement. At this time, the CT chest revealed extensive ground glass opacities with relative sparing of the lung bases, suggestive of diffuse pulmonary haemorrhage. He required an 8-day ICU admission for vasopressor support but did not require intubation. He made a full recovery and was discharged from hospital after an 11-day admission.

3.2. Pulmonary Haemorrhage

Pulmonary haemorrhage was diagnosed in 26/109 (24%). Individuals with pulmonary haemorrhage were more likely to be current smokers than individuals without pulmonary haemorrhage (14/26 (54%) versus 26/83 (31%), p=0.04). There were no other demographic or clinical associations with the development of pulmonary haemorrhage. In particular, there was no association with the development of pulmonary haemorrhage and the presence of underlying lung disease or the infecting serovar. The median (IQR) duration of symptoms prior to the initiation of antibiotic therapy in the individuals who developed pulmonary haemorrhage was 5 (4-6) days versus 4 (3-5) days in the individuals who did not develop pulmonary haemorrhage, but this difference did not reach statistical significance (p=0.10) (Table S7).

Among the 26 individuals with pulmonary haemorrhage, 16 (62%) had respiratory symptoms at presentation: 14 (54%) had a cough, 12 (46%) had haemoptysis and 8 (31%) complained of dyspnoea; 18 (69%) had abnormal chest auscultation at presentation. Individuals with pulmonary haemorrhage had lower SpO₂/FiO₂ ratios at presentation and were more likely to have oliguria and require vasopressor support than patients who did not have pulmonary haemorrhage. The other clinical and laboratory findings in the individuals who developed pulmonary haemorrhage are presented in Tables S8, S9 and S10. At presentation, individuals who developed pulmonary haemorrhage had greater renal impairment, a higher C-reactive protein and were more likely to have an abnormal troponin than individuals without lung involvement; 24/25 (96%) individuals with pulmonary haemorrhage who had PCR testing of blood were PCR positive and 7/15 (50%) who had blood cultures collected had a positive culture.

3.4. Clinical Course and Therapy

It was not possible to confirm the diagnosis of leptospirosis for several days and prior to this, patients tended to receive standard ward-based management guided by the patients’ physiological parameters. This care included antibiotic therapy, cautious crystalloid fluid therapy, and prompt initiation of additional supportive care – including vasopressors, RRT mechanical ventilation and ECMO – as necessary, using standard indications [31].

Individuals with lung involvement frequently had multi-organ dysfunction and were more likely to require ICU admission than individuals without lung involvement (41/62 (66%) versus 15/47 (32%), odds ratio (95% confidence interval): 4.17 (1.86-9.34), p<0.001) (Figure 3 and Table 4). Arterial blood gas analysis was performed in 41/62 (50%) individuals with lung involvement; the median (IQR) nadir PaO₂/FiO₂ ratio was 236 (151-303); 30/62 (48%) satisfied the Berlin criteria for ARDS: 12/30 (40%) had mild, 15/30 (50%) had moderate, and 3/30 (10%) had severe ARDS. Intubation and mechanical ventilation were necessary in 15/62 (24%); all 15 of these individuals had lung involvement. The demographic, clinical and laboratory indices in individuals requiring mechanical ventilation are presented in Tables S11, S12 and S13.

There were 10/62 (16%) who received vasopressor support, RRT and mechanical ventilation; 5/62 (8%) who received vasopressor support and mechanical ventilation; 3/62 (5%) who received vasopressor support and RRT; 23/62 (37%) who received only vasopressor support and 2/62 (3%) who received only RRT. There were 19/62 (31%) individuals with lung involvement who did not receive vasopressor support, RRT or mechanical ventilation.

Of the 109 patients, 107 (98%) received antibiotic therapy at some point in their hospitalisation. One patient (a patient with lung involvement) had a Jarisch-Herxheimer reaction, but this was managed successfully with simple supportive measures. A range of antibiotics were prescribed for a variety of durations, but doxycycline (81/109 (74%)) and ceftriaxone (74/109 (68%)) were the agents prescribed most commonly. The two individuals who did not receive antibiotics were children (aged 7 and 11 respectively) who had the diagnosis of leptospirosis confirmed after the resolution of symptoms; neither had lung involvement.

Corticosteroids were prescribed in 19/109 (17%) and were prescribed more often in individuals with lung involvement (16/62 (26%) versus 3/47 (6%), p=0.01). However, this corticosteroid therapy was usually intravenous hydrocortisone (typically at a dose of 50mg four times a day) in the setting of refractory hypotension in individuals requiring vasopressor support; all 19 patients who received steroids received vasopressor support. Two individuals received high dose pulse methylprednisolone for severe pneumonitis, while one patient received dexamethasone in the setting of concurrent severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. The median (IQR) duration of steroid therapy was 2 (1-4) days.

Two patients with refractory respiratory failure required evacuation to a quaternary centre to receive ECMO; 1 of these individuals also received intravenous tranexamic acid and nebulised adrenaline. Inhaled tranexamic acid was prescribed to 3 of the other 25 individuals with pulmonary haemorrhage. No patients received plasma exchange or plasmapheresis. All patients in the cohort were alive 90 days after presentation.

4. Discussion

The lungs are commonly affected in patients hospitalised with leptospirosis in tropical Australia, with radiological evidence of lung involvement identified in more than half of the patients in this cohort. This lung involvement is life-threatening: almost half of the patients with lung involvement satisfied criteria for ARDS and over 40% had pulmonary haemorrhage. Almost two-thirds of the patients with lung involvement required ICU admission, although importantly, lung involvement was commonly just one element of multi-organ dysfunction in these patients.

However, although leptospirosis was considered in the initial differential diagnosis of approximately 60% of the patients with lung involvement, the diagnosis could only be confirmed several days into the admission. Prior to this, patients tended to receive standard sepsis management which included antibiotic therapy, cautious fluid resuscitation with crystalloid, vasopressors (with low-dose corticosteroids as necessary) to maintain blood pressure and prompt initiation of additional supportive care – including mechanical ventilation, RRT and ECMO – guided by the patients’ physiological parameters and clinical trajectory [31]. Fortunately, with access to care in Australia’s well-resourced health system, every patient in the cohort – including the 57% with confirmed lung involvement – was alive 90 days after presentation.

Over 95% of the patients with lung involvement had no history of chronic lung disease. Indeed, over 40% of the patients with lung involvement were aged younger than 40 and had no comorbidity, highlighting the pathogen’s ability to cause life-threatening disease in otherwise well young people. The minority of patients with lung involvement during their illness had respiratory symptoms at presentation, only 55% had an abnormal respiratory examination and only 58% had abnormal chest imaging, emphasising the importance of frequent clinical reassessment and a low threshold for repeat chest imaging in patients in whom leptospirosis is suspected, as clinical deterioration can be rapid [32].

The imaging findings in our cohort were similar to those identified in other series; multilobar and alveolar changes, in particular, were common and were associated with a more complicated disease course [33,34,35]. These radiological appearances would correlate with the histological findings of alveolar oedema and haemorrhage that are seen commonly in patients with fatal leptospirosis [36,37,38,39]. The precise pathophysiology of lung involvement in patients with leptospirosis remains incompletely understood, but human autopsy studies have demonstrated leptospiral antigen within endothelial cells of the interalveolar septa and attached to capillary endothelial cells, supporting the hypothesis that the organism exerts a locally destructive action on lung capillaries, leading to loss of vascular integrity and the alveolar oedema and haemorrhage which is seen in severe disease [32,37,38,40].

These clinicopathological correlation studies provide support for the management strategies employed in the patients with lung involvement in our cohort. Antibiotic therapy was prescribed to all but two individuals, both children without lung involvement. Some authors have expressed doubts about the incremental value of antibiotic therapy in patients with leptospirosis presenting later in their disease course and a Cochrane review – which emphasised that few quality studies have examined the issue – found little evidence to suggest benefit of antibiotic therapy in patients with the infection [5,12]. However, although patients with lung involvement in this cohort presented after a median of 5 days of symptoms, leptospiraemia was documented in 95% of those who had blood PCR performed or blood cultures collected (and in every patient with lung involvement who required ICU care). Indeed, the development of lung involvement was associated with delays in the administration of antibiotic therapy, echoing findings from other series and providing data to support the utility of antibiotic therapy in these patients [41,42].

In addition to the human autopsy studies demonstrating alveolar oedema – hypothesised to result from direct vascular injury – animal models of leptospirosis have documented decreased epithelial sodium channel (H-ENaC) protein expression and upregulation of the Na-K-2Cl cotransporter NKCC1 in the lungs which further impairs pulmonary fluid handling [36,37,38,43]. Although hypotension and AKI were common in the cohort, concerns about precipitating or exacerbating pulmonary oedema meant that fluid resuscitation was generally conservative in these patients, even before changes were apparent on chest imaging. Large fluid boluses were avoided in patients with hypotension or AKI who were instead managed using standard sepsis algorithms that included prompt vasopressor support to maintain a mean arterial pressure and, where necessary, RRT [31].

Pulmonary haemorrhage is the most common cause of death in patients with leptospirosis and even in ICUs with access to ECMO mortality can exceed 30% [19,38,39,44,45,46]. Pulmonary haemorrhage was identified in 26 individuals in our cohort and although this necessitated ECMO in two cases – and three others received tranexamic acid – the patients’ management was generally supportive and expectant, and outcomes were excellent. Although plasmapheresis is available locally, no patient had plasmapheresis or plasma exchange. There has been uncertainty about the contribution of disseminated intravascular coagulation (DIC) to the development of pulmonary haemorrhage, however while patients with pulmonary haemorrhage had a lower nadir platelet count than patients without pulmonary haemorrhage, their coagulation studies did not suggest the presence of DIC [47]. It was notable that pulmonary haemorrhage was more common in current cigarette smokers, echoing findings from a New Caledonian series [42]. The explanation for this possible association is uncertain although it is plausible that smoking could lead to damage to the alveolar basement membrane, increasing the risk of haemorrhage [48,49]. The fact that over a third of the cohort were current cigarette smokers – compared to a current rate of 8.3% in Australians aged 14 or older – may have contributed to the high incidence of pulmonary haemorrhage in our study [50].

Many authors have suggested that a dysregulated and intense immune response plays an important role in the pathogenesis of severe disease [5,51,52]. This is the basis for the high dose corticosteroid therapy that has been proposed by some authors, particularly for patients with lung involvement [53]. However, while infiltration of alveolar spaces by monocytes and neutrophils occurs, inflammatory infiltrates are generally not prominent [37,38,39,54]. Although almost 85% of the patients in our cohort with lung involvement received corticosteroids, this was usually modestly dosed hydrocortisone that was prescribed for less than 72 hours in the setting of hypotension requiring vasopressor support, rather than the high dose bolus methylprednisolone for three days (followed by oral corticosteroids for five-seven days) that has been used in some series [16,53]. The encouraging outcomes seen in our cohort suggest that higher dose corticosteroid therapy should not be considered as a routine intervention in patients with lung involvement, particularly given the risk of nosocomial infection, until definitive data from adequately powered randomised controlled trials is available to support this approach [55,56].

It is important to highlight that even in Australia’s well-resourced health system where there is access to a World Health Organization Collaborating Centre for Leptospirosis, the diagnosis of leptospirosis in our cohort was frequently not confirmed for 5-7 days, and in some cases it was a retrospective diagnosis [28]. The clinical and laboratory findings in patients with leptospirosis, while frequently suggestive, are often non-specific and it can be difficult to distinguish patients with leptospirosis from patients with other life-threatening infections and, indeed, non-communicable diseases [31,47,57,58,59]. Therefore, it is important not to institute specific interventions for the management of leptospirosis that fails to consider these other diagnoses or contains elements that may – either by omission or commission – increase the risk of deterioration. In this context the pragmatic clinician can be reassured that many elements of the care received by the patients with laboratory-confirmed leptospirosis in this cohort would be equally appropriate in patients with the other infectious diseases that may mimic it. These include early antibiotic therapy [31,60,61], a conservative approach to fluid resuscitation [62,63,64,65], cautious use of corticosteroids [66,67], and the early delivery of context dependent critical care support [31,68,69].

Although this study was able to examine the demographic, clinical, laboratory and radiological findings of patients with lung involvement complicating laboratory-confirmed leptospirosis in some detail, it has several limitations. The study’s retrospective nature meant that patients’ clinical assessment, investigations and management were not standardised. They occurred at different stages in the evolution of the disease and the reporting of comorbidities and clinical findings, relied on accurate documentation in the medical record. The study examined only patients managed in a referral hospital and therefore included patients with a more severe clinical phenotype, overestimating the incidence of severe disease. The definition of lung involvement, while pragmatic, is imperfect. Although we used the imaging reports of a specialist radiologist, a plain chest x-ray may not identify early parenchymal changes and lung imaging was performed only once in a third of the cohort, which will tend to underestimate the incidence of lung involvement [35]. The categorical definition of lung involvement lacked nuance; there is clearly a significant clinical difference between subtle, asymptomatic chest x-ray changes and multilobar alveolar changes causing respiratory failure. The definition of pulmonary haemorrhage was equally crude and required adequate documentation in the medical record. There is some geographical variation in the clinical presentation of leptospirosis which some clinicians might suggest limits the applicability of our observations to other settings [70,71,72]. Infection with serovars from the Icterohemorrhagiae serogroup has been reported to be associated with lung involvement and pulmonary haemorrhage [32,72,73,74]. However, it was notable that serovars from the Pyrogenes serogroup (Zanoni and Robinsoni) and Australis serogroup (Australis), the most common serovars in the region, were also the most commonly identified serovars in patients with lung involvement and pulmonary haemorrhage in this cohort. These data support the contention that there may be limited correlation between particular serovars and the severity and clinical manifestations of leptospirosis [51,75,76].

It is important to highlight that the patients in our cohort were managed in Australia’s well-resourced, universal health system which limits the generalisability of some of our recommendations, particularly to the resource-limited settings where, globally, most cases of leptospirosis are seen [1]. That said, we would argue that the clinical approach and management strategies which include prompt antibiotic therapy, conservative fluid resuscitation, efforts to exclude alternative diagnoses and the provision of additional supportive care – where this is available – are likely to be equally relevant in these locations [31,69]. Although we have presented the general approach to the management of patients with leptospirosis in the FNQ region, detailed quantitative data describing the different interventions (including precise volumes of fluid, ventilation strategies, blood product support and the indications for mechanical ventilation, RRT and ECMO) have not been presented. They will be the focus of a manuscript that is in preparation. Separate reports will also examine the manifestations of leptospirosis in other organ systems in patients in the FNQ region.

5. Conclusions

Life-threatening lung involvement can complicate leptospirosis in young and otherwise well individuals. However, in Australia’s well-resourced health system excellent outcomes can be achieved using a standard contemporary approach to patient management while a confirmed diagnosis is awaited.