Introduction

Leptospirosis is a globally prevalent zoonotic disease caused by Leptospira spp., Gram-negative spirochetes that typically use mammals—mainly rodents—as their natural reservoirs [1]. In tropical continental countries such as Brazil, leptospirosis can exhibit high incidence rates, particularly in endemic areas during the rainy season and following flooding events. In the Northeast region, between 2019 and 2022, 2,199 cases of leptospirosis were reported, with a case fatality rate of 13.8% [2]. The bacteria colonize the renal tubules of reservoir hosts and are excreted in their urine, contaminating the environment and enabling transmission to humans, who are considered incidental hosts [1]. The incubation period in humans is variable, ranging from one day to four weeks after exposure, and in survivors, the infection may persist for several months [3].

After contamination, the spirochetes disseminate through the bloodstream during an approximately eight-day period known as the leptospiremic phase, in which a higher pathogen load may result in more severe disease [1]. The subsequent immune phase corresponds to the immunological response to infection [4]. Clinically, leptospirosis can range from a mild febrile illness to a severe icteric–hemorrhagic syndrome with multiple organ dysfunction [5], first described by Adolph Weil in 1886 [6]. Despite extensive research, important gaps remain in understanding the pathogenesis of this tropical disease, particularly regarding the extent to which clinical manifestations result from direct bacterial invasion versus host immune-mediated injury [1].

Although it is described and frequently cited as part of the disease spectrum [7], leptospirosis-associated meningoencephalitis (LAM) remains poorly characterized in the medical literature. Most publications addressing this topic were released before the 21st century, often in low-impact scientific journals, and many of these reports are not available online, hindering the consolidation of current knowledge on LAM. Several factors may contribute to this gap: (i) LAM can clinically mimic other causes of aseptic meningoencephalitis, such as viral, mycobacterial or rickettsial infections [8]; (ii) appropriate diagnostic tools for central nervous system (CNS) involvement are often unavailable in regions where leptospirosis is endemic; and (iii) the empirical use of antibiotics during the course of meningoencephalitis may reduce the likelihood of isolating the pathogen [9].

Consequently, establishing a consistent clinical–epidemiological–laboratory profile of LAM remains challenging, as it may encompass a wide range of neurological syndromes, including meningitis [10], meningoencephalitis [9], acute encephalitis [11], demyelinating diseases [12,13], cerebrovascular disorders [14], focal CNS lesions [15], peripheral neuropathies [16], and Guillain–Barré syndrome [17], among others. In addition, important inconsistencies persist across current studies, largely driven by methodological limitations, including small sample sizes, reliance on isolated case reports, heterogeneous diagnostic criteria, and variability in laboratory methods, which collectively hinder meaningful comparisons and synthesis of findings. For example, some reports describe LAM exclusively in anicteric leptospirosis, whereas others demonstrate its occurrence in severe forms such as Weil’s disease [18,19].

We characterized LAM cases among patients with aseptic meningitis at a tertiary infectious diseases center in Ceará, Northeastern Brazil, and performed a literature review and meta-summary of case reports, series, and cohorts to better define its clinical and cerebrospinal fluid (CSF) profiles.

Methods

Results

Cases description

Case 1. A 22-year-old Brazilian man who works in an urban area as a motorcycle delivery worker, and without previous comorbidities, arrived at the emergency room (ER) with a 7-day history of high fever (Temperature 40°C), headache, nuchal stiffness, polyarthralgia, macroscopic hematuria and nausea. He reported prolonged occupational exposure (>1 hour) to stagnant rainwater one week prior to admission.

Vital signs included a heart rate of 94 bpm, a respiratory rate of 20 rpm, and a peripheral oxygen saturation (SpO₂) of 99% without supplementary oxygen. On physical examination, he presented with mild jaundice and a pain-related facial expression. Neurological examination revealed positive Kernig’s and Brudzinski’s signs, with a negative Lasègue sign. He was alert and oriented (Glasgow Coma Scale score of 15). Laboratory tests showed hyperbilirubinemia, thrombocytopenia, and renal dysfunction. Cranial CT was unremarkable. CSF analysis was consistent with aseptic meningitis (Table 1). Serum anti-Leptospira IgM ELISA (Panbio, Abbott, Australia) was positive, and microscopic agglutination testing (MAT) showed a titer of 1:400. The serovar was not identified.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Cerebrospinal fluid profile of patients with severe leptospirosis-associated meningitis.

https://doi.org/10.1371/journal.pntd.0014315.t001

Empirical treatment for bacterial meningitis with intravenous ceftriaxone (2 g, twice daily), vancomycin (15 mg/kg three times daily), and intravenous dexamethasone (12 mg, three times daily) was initially performed; vancomycin and dexamethasone were discontinued after leptospirosis confirmation. The patient improved clinically, although he developed acute pancreatitis during hospitalization, managed conservatively. He achieved full recovery and was discharged after four weeks.

Case 2. A 66-year-old Brazilian man, a construction worker with hypertension and alcohol use, presented with a 2-week history of fever, headache, polyarthralgia, disorientation, and sensory deficits. He reported a recent lower-limb laceration sustained during flood exposure.

Vital signs included a heart rate of 124 bpm, a temperature of 37,8ºC, a respiratory rate of 16 rpm, and a peripheral oxygen saturation (SpO₂) of 88% without supplementary oxygen. On physical examination, Glasgow Coma Scale score of 11, with marked sensory loss. Neurological examination revealed nuchal stiffness. Laboratory tests showed anemia, leukocytosis, elevated inflammatory markers, hyperbilirubinemia, and mild renal dysfunction. CSF findings were consistent with aseptic meningitis (Table 1). Serum anti-Leptospira IgM ELISA was positive (Panbio, Abbott, Australia). MAT was unavailable at the time. returned a positive result.

The patient received ceftriaxone and vancomycin and required ICU admission with mechanical ventilation due to neurological impairment. He improved clinically, with recovery of renal function and no need for dialysis. Other complications, such as alveolar hemorrhage, were investigated through thoracic CT and ruled out. After 14 days of antimicrobial therapy, he was discharged without neurological sequelae, confirmed at 6-month follow-up.

Case 3. A 25-year-old transgender woman, a waste picker with newly diagnosed type 2 diabetes, presented with myalgia, calf pain, neck stiffness, and a recent convulsive episode. She reported occupational exposure to stagnant rainwater.

On admission, she was hemodynamically stable (temperature 37.8°C) and fully alert (GCS 15), with no focal neurological deficits. Laboratory evaluation revealed severe rhabdomyolysis (CPK 72,000 U/L) and acute kidney injury (creatinine 3.5 mg/dL, urea 115 mg/dL). Cranial CT was normal. CSF analysis demonstrated lymphocytic meningoencephalitis with low cellularity and protein (Table 1). Serum anti-Leptospira IgM ELISA (Panbio, Abbott, Australia) was positive.

She was treated with intravenous ceftriaxone and supportive care. Renal dysfunction was managed conservatively without dialysis. The patient remained neurologically stable, with no recurrence of seizures. She was discharged with partial renal recovery and outpatient follow-up.

Literature review

In this study, we present three rare cases of a common disease - leptospirosis. Given the uncertainties regarding the possible axial and extra-axial manifestations of the disease, we conducted a literature review with a meta-summary of published case reports. Through this approach, we identified 44 articles describing 179 cases with available clinical and CSF profiles.

The retrieved articles were summarized in Table 2. The aggregated data show that, among the studies with available information on jaundice, the majority of patients with leptospiral meningitis/meningoencephalitis were anicteric (59.1%), while 40.9% presented with jaundice. Four studies did not express the prevalence of icteric/anicteric patients in their cohort [8,23–25].

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Studies and cases on leptospirosis-associated meningitis in medical literature.

https://doi.org/10.1371/journal.pntd.0014315.t002

The articles that provided CSF data were allocated to a spreadsheet; total missing data was 44.1% (Fig 1). A total of 179 patients had their CSF analyzed. A lymphocytic inflammatory pattern was described, with a median low-to-moderate cellularity (approximately 68 cells/mm³) and wide variation, including a few very high values (Fig 2). The lymphocyte ratio is high (median close to 73%), while neutrophils remain low (median close to 18%). However, among the 42 patients for whom CSF neutrophil data were available, 12 (28.6%) demonstrated a predominance of neutrophils in the CSF. CSF glucose tends to be within the normal range (median ~60 mg/dL), and protein shows a slight to moderate elevation (median ~67–90 mg/dL), although with rare very high values. Taken together, the findings reinforce a typical profile of aseptic meningitis, with lymphocytic predominance, absent hypoglycorrhachia, and variable proteinorrhachia.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Heatmap summarizing missing data patterns across CSF parameters from literature and the three reported cases.

The dataset combines variables extracted from published studies reporting CSF findings in leptospirosis-associated meningitis, together with the three additional cases described in the present work. Studies included: Z-Y.

https://doi.org/10.1371/journal.pntd.0014315.g001

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Boxplot of CSF parameters from literature and the three reported cases.

Extremely high outlier values that distorted the graphical visualization were removed prior to plotting (protein > 400 mg/dL and glucose > 150 mg/dL). Studies included: Z-Y.

https://doi.org/10.1371/journal.pntd.0014315.g002

Two reports were excluded from Fig 2 due to markedly elevated protein levels in the CSF, measuring 875 mg/dL and 2590 mg/dL, respectively [15,39]. Additionally, two cases exceeded the predefined upper threshold for CSF cellularity, with 7800 cells/mm³ and 5000 cells/mm³, and were therefore also removed from the analysis [39].

Discussion

The present study comprises a retrospective cohort of aseptic meningitis cases managed at a tertiary infectious diseases center in northeastern Brazil, with identification and characterization of confirmed cases of LAM.

In our cohort, 3/447 (0.67%) of patients with aseptic meningitis had serological evidence compatible with leptospirosis, including one case of isolated meningitis and two cases of meningoencephalitis, all of which exhibited lymphocytic predominance in the CSF. In contrast, Romero et al. (2010) [40] reported a 58.97% positivity rate for Leptospira spp. using molecular methods in CSF samples from patients with aseptic meningitis, a proportion substantially higher than that observed in our study. This discrepancy may be explained by methodological differences, as PCR-based techniques tend to demonstrate greater sensitivity [8]. In another study conducted in the same region, IgM ELISA positivity in CSF was observed in 14.6% of cases of meningitis of indeterminate etiology [43]. In a large series of leptospirosis, meningeal involvement has been reported in up to 37% of patients, and CSF pleocytosis in up to 47% [56]. Therefore, the likelihood that test positivity in the context of meningitis merely reflects incidental exposure to Leptospira is considered low [9].

Given the scarcity of detailed descriptions of this condition in the literature, a metasummary of reported cases was conducted to delineate the clinical and CSF profile of patients with meningeal syndromes associated with leptospirosis (Fig 2 and Table 2). Overall, a predominantly lymphocytic profile was observed, with mildly increased cellularity, normal to elevated CSF glucose levels, and marked elevation of CSF protein (Fig 2). Nevertheless, substantial heterogeneity in CSF findings was evident, with some cases demonstrating neutrophilic predominance—several of which were confirmed by molecular methods in the CSF [9,18,39,26,32,36,38,51], as well as others showing markedly elevated protein and cell counts [39,32,36,48,49]. These findings underscore the broad spectrum of CSF abnormalities associated with CNS leptospirosis and highlight the diagnostic challenges inherent to this condition.

In milder forms of the disease, it remains unclear to what extent neurological manifestations result from direct pathogen-mediated injury or from host immune responses [1,57,58]. Similar to the systemic infection, CNS involvement is thought to follow a biphasic course [1,58]. There is substantial evidence that, after penetrating host barriers, pathogenic Leptospira rapidly disseminates hematogenously, leading to the acute leptospiremic phase [58]. During this stage, the organism can be isolated from multiple tissues, including the CNS, often without eliciting significant inflammatory responses, which explains the absence of pleocytosis or overt inflammatory changes in some cases [58,59]. Clinically, this phenomenon is supported by studies demonstrating molecular evidence of Leptospira in the CNS despite unremarkable inflammatory parameters [39]. In contrast, the immune phase of the disease is believed to account for the majority of neurological manifestations, including meningitis, meningoencephalitis, and encephalitis, driven by antigen–antibody–mediated hypersensitivity reactions [59]. Although uncommon, these cases often demonstrate CSF positivity by serological methods such as MAT and ELISA [40–42,44–47].

Herein, we describe three cases of LAM with acute presentation, in accordance with the majority of published reports on LAM. Counterintuitively, the study by Murgatroyd (1937) described a case of chronic LAM with compatible clinical and CSF findings in which Leptospira was isolated from CNS by xenodiagnosis [50]. This observation demonstrates that leptospirosis may present with CSF pleocytosis and active central nervous system infection, rather than representing solely an immune-mediated phenomenon [50]. In that case, the patient exhibited immunological anergy to their infecting Leptospira strain, requiring experimental administration of equine antileptospiral serum for disease control, suggesting that chronicity resulted from either low bacterial virulence or impaired host immune response [50]. In a convergent manner, Wilson et al. (2014) reported a case of LAM in an adolescent with a four-month history of meningoencephalitis who had a previous diagnosis of severe combined immunodeficiency [27]. This finding supports the hypothesis that persistence of Leptospira infection may occur when pathogen clearance after the bacteremic phase is impaired, whether due to host immune deficiency or intrinsic characteristics of the infecting strain. None of our three patients were immunodepressed, and only these two studies reported impairment of immune response among the LAM patients. Thus, isolation of the organism may still occur during the so-called “immune phase.” Regardless of complement system integrity, effective immune control of leptospirosis relies on the generation of functional, pathogen-specific antibodies. Host-related factors—including comorbidities, immunosuppression, and genetic polymorphisms—play a critical role in shaping the clinical spectrum of disease [1]. Other studies likewise support the possibility of clinically apparent CNS infection during the bacteremic phase of leptospirosis [18,45].

The pathogenesis of LAM remains poorly understood, largely owing to the frequent underdiagnosis of this condition. Molecular-based prevalence studies conducted in endemic areas generally report higher detection rates, suggesting that CNS involvement may be more common than clinically recognized [40]. The frequent use of antibiotics in cases of acute meningitis or systemic leptospirosis contributes substantially to this limitation [39,57]. Nevertheless, experimental studies and clinicopathological reports of CNS infection by Leptospira spp. provide important insights into the underlying mechanisms. The immunological nature of LAM is further supported by one study, in which the patient developed meningeal symptoms only after the initiation of antibiotic therapy, consistent with a Jarisch–Herxheimer reaction [38]. This phenomenon, characterized by the release of bacterial cell wall components and subsequent inflammatory cytokine surge, is well described in infections such as leptospirosis and syphilis [60].

Histopathological data remain scarce; however, findings of the studies retrieved in Table 2 reveal cerebral edema and congestion, along with mixed inflammatory infiltrates composed of histiocytes, lymphocytes, and plasma cells [35]. Microglial nodules were also observed across multiple brain regions. Additional findings included perivascular hemorrhage and demyelination, accompanied by varying degrees of lymphohistiocytic infiltration and vascular leakage, consistent with a post-infectious acute disseminated encephalomyelitis–like process without spinal cord involvement. Other studies corroborate these observations, demonstrating predominantly perivascular lymphohistiocytic infiltration within the periarachnoid space [27]. In line with neuropathological findings, LAM is strongly associated with cerebrovascular involvement. Our three cases did not present with CNS bleeding or infarction, but several studies of the meta-summary reported brain vasculopathy (Table 2). Although the exact mechanisms remain unclear, it is widely accepted that immune-mediated vasculitis and/or direct vascular invasion play a central role, contributing not only to intracranial hemorrhage but also to cranial nerve involvement, both of which are frequently observed in this condition [30,45,61]. Additionally, Leptospira spp. can interfere with the fibrinolytic system, coagulation pathways, platelet function, and endothelial activation, thereby increasing the risk of hemorrhagic complications [62]. A population-based cohort study from Taiwan demonstrated an increased risk of hemorrhagic stroke among patients with newly diagnosed leptospirosis, even in the absence of overt meningeal involvement [63]. In the reported cases of LAM, cerebrovascular manifestations may include lacunar infarction [13,30,35], subarachnoid hemorrhage [31,45], territorial ischemic stroke [35], Moyamoya-like vasculopathy [8,14], extra-axial hemorrhage [64], and cerebral venous thrombosis [65], among other presentations.

Molecular studies in LAM have expanded understanding of both diagnostic strategies and disease pathophysiology [8,13,27,40,51,66,67]. Evidence suggests that CSF PCR is more sensitive than serological assays, enabling earlier diagnosis and therapeutic intervention [40]. However, its use remains limited in many endemic regions, contributing to the underrecognition of LAM in patients presenting with aseptic meningitis [15,40]. Importantly, the detection of Leptospira DNA in patients with LAM does not necessarily indicate active infection, as it may reflect residual genetic material rather than viable organisms capable of replication [15]. These findings suggest that leptospiral aseptic meningitis is likely substantially underreported as a cause of meningitis, particularly in settings where molecular diagnostic tools are not routinely available. Prior studies incorporating PCR-based approaches have demonstrated higher detection rates, reinforcing the notion that the true burden of LAM may be underestimated in the existing literature [66,67].

This study has limitations that should be acknowledged. First, the diagnosis of LAM was not confirmed by direct detection of Leptospira in CSF, which may limit diagnostic certainty. However, all cases met accepted diagnostic criteria based on compatible clinical and epidemiological features in conjunction with positive serology, supporting the validity of the diagnoses. Second, the available literature on LAM is scarce and frequently characterized by incomplete reporting and substantial heterogeneity in clinical, laboratory, and diagnostic data, which constrains comparability and synthesis. Third, the retrospective design of this study may be subject to information bias and limits the ability to establish temporal and causal relationships. Finally, it was not possible to further characterize the serovars involved in the studies, as most diagnoses were based on serology/MAT, which cannot reliably identify the infecting serovar [28,66,67].

An estimated one million cases of leptospirosis occur worldwide each year, accounting for approximately 60,000 deaths [1]. Even in its classic clinical presentations, the disease is frequently underrecognized; atypical manifestations further complicate diagnosis and may contribute to substantial underdiagnosis. Treatment of leptospirosis should be initiated based on clinical suspicion, as early antimicrobial therapy is essential regardless of whether neurological involvement results from immune-mediated mechanisms or direct bacterial invasion. Nevertheless, these findings contribute meaningfully to the understanding of CNS involvement in this neglected tropical disease.

Conclusion

Leptospiral involvement of the CNS remains an underrecognized and likely underreported cause of aseptic meningitis, as reflected by its low frequency in our cohort despite occurring in an endemic setting. The marked variability in CSF findings underscores the diagnostic complexity of this condition and limits reliance on classical laboratory patterns. Differences in detection rates across studies highlight the importance of sensitive molecular techniques, which may reveal a higher true burden of disease. Current evidence suggests a multifactorial pathogenesis, involving both direct bacterial invasion and immune-mediated mechanisms, with host factors influencing clinical expression and disease persistence. From a clinical perspective, LAM should be considered in the differential diagnosis of aseptic meningitis in endemic areas, particularly in the presence of relevant epidemiological exposures and clinical signs of leptospirosis, as early recognition and timely antimicrobial therapy may improve outcomes. From a research standpoint, expanding access to molecular diagnostics and conducting well-designed prospective studies are essential to better characterize disease burden, elucidate underlying mechanisms, and guide evidence-based diagnostic and therapeutic strategies for leptospiral CNS involvement.

Acknowledgments

The authors gratefully acknowledge the institutional support provided by the Universidade de Fortaleza (UNIFOR), particularly the Postgraduate Program in Collective Health and Medical Sciences, for fostering the development of this research. The authors also thank UNIFOR for its continued commitment to academic excellence and support for clinical and epidemiological research in tropical diseases. We also would like to acknowledge Christus University Center and São José Hospital for Infectious Diseases for their valuable support during the research and data collection.