Clinical complexities in diagnosing and managing leptospirosis in New Zealand: a qualitative analysis

Leptospirosis is a globally under-diagnosed neglected zoonotic disease that presents as a mild to fulminant undifferentiated febrile illness, which can be fatal.^1,2 The World Health Organization (WHO) guidelines include four broad clinical categories: 1) mild, influenza-like illness; 2) Weil’s syndrome, characterised by jaundice, renal failure, haemorrhage and myocarditis with arrhythmias; 3) meningitis/meningoencephalitis; and 4) pulmonary haemorrhage with respiratory failure.³ The New Zealand Ministry of Health – Manatū Hauora describes leptospirosis as an acute illness characterised by fever, chills, headache, myalgia, nausea, diarrhoea, abdominal pain, meningitis, cough and conjunctival suffusion; manifestations of severe disease can include jaundice, renal failure, haemorrhage, pneumonitis and haemodynamic collapse.⁴ Leptospirosis’s clinical similarity to other infectious diseases makes diagnosis challenging, a difficulty further compounded by the biphasic nature of the illness and the different diagnostic tests, which vary in sensitivity depending on the disease stage.^2,5,6 In the acute stage (leptospiraemia, 1–7 days post–disease onset), diagnosis primarily relies on PCR testing of blood and cerebrospinal fluid, whereas in the convalescent stage (immune phase, >7 days), diagnosis relies on serology and polymerase chain reaction (PCR) testing of urine.² If the recommended test is not used at the appropriate stage of the disease, it may result in false negatives.

In New Zealand, 53% of notified leptospirosis cases between 1999 and 2017 were hospitalised.⁷ Although the disease can be treated with antibiotics, evidence suggests that some patients experience long-term sequelae following infection, highlighting both the acute and long-term burden.⁸ Earl et al. demonstrated that one in four patients in New Zealand with flu-like symptoms and a relevant exposure history tested positive for leptospirosis.⁹ Similarly, Irwin et al. identified that leptospirosis is likely under-diagnosed in routine clinical practice in New Zealand.¹⁰ These studies highlight that although leptospirosis is often perceived as uncommon, it may be more frequently encountered than clinicians expect.

To date, few studies have examined the clinical reasoning processes influencing diagnostic decisions for leptospirosis. One study in Thailand explored clinicians’ use of laboratory parameters,¹¹ and one in New Zealand described clinical presentation and severity through a review of hospital records.¹² Gaining insight into clinicians’ diagnostic reasoning may help identify opportunities to enhance early detection and management. This study analysed clinical notes of diagnosed leptospirosis cases to identify patterns and themes, explore their implications and develop recommendations to improve the diagnosis and management of leptospirosis in New Zealand.

Methods

Case selection

Leptospirosis cases in this study included a subset of cases (42 of 95) from 2019 to 2022 that were part of a nationwide case-control study.¹³ Cases were selected if they had a clinically compatible illness with one or more of the following laboratory evidence: 1) isolation of leptospires from a clinical specimen, 2) detection of leptospiral DNA from a clinical specimen using PCR, 3) detection of a ≥4-fold rise in serology titres between acute and convalescent sera using the microscopic agglutination test (MAT), 4) detection of a single raised titre using MAT, and 5) immunoglobulin M–positive or equivocal result. Full details of case selection were published in a protocol paper.¹⁴

Data source

This study acquired clinical notes of 42 leptospirosis cases from the general practitioners (GP) of the cases. Clinical notes were securely shared with the investigators and anonymised before analysis. Notes included GP visits, phone consultations in primary care, referrals to secondary services, referrals between secondary services, discharge summaries from the emergency department and inpatient medical teams, outpatient letters and diagnostic test results from primary and secondary services. Data were drawn from initial presentation to last follow-up for the episode of leptospirosis.

Data analysis

This study used qualitative content analysis of clinical notes from leptospirosis cases to identify key themes using a dual deductive and inductive approach.¹⁵ The deductive analysis organised clinical records into categories based on a medical consultation framework including clinical history and exposure assessment; presenting symptoms and signs; formulation of differential diagnoses; ordering and interpretation of diagnostic tests and treatment, management and follow-up plans.

The inductive analysis focussed on summarising each patient’s diagnostic and treatment journey using narrative text. These narratives were coded using a constant comparison method to identify recurrent patterns and thematic issues across cases. Findings from the deductive analysis were integrated into inductive analysis to explore the factors influencing decision making in practice. Reflections and discrepancies were documented throughout the analysis and informed the theme development, illustrated by participant quotes.

Coding patterns and themes were developed by the first author (a rural GP with an understanding of complexity in general practice) and the last author (an epidemiologist with expertise in zoonotic disease) through a consensus approach to build interpretive rigor.

Ethics approval

This study received human ethics approval from the Health and Disability Ethics Committee (reference number 19/STH/80) as well as locality agreements and local Māori consultations from 20 district health boards.

Results

Clinical data varied widely, reflecting the unique journey of each patient’s illness and the diversity of clinical processes. Individual records ranged from 0 to 31 pages, including letters up to 400 words and laboratory results up to 16 pages. Of the 42 clinical notes reviewed, 29 provided biochemistry and haematology results, 32 provided diagnostic serology and/or PCR results, 35 provided information on presenting symptoms and signs, 34 had text on differential diagnosis development, 34 contained management information and 33 contained confirmatory testing information. Three sets of notes contained no information, and five cases had no data on treatment. Approximately 60% (25/42) of the cases were hospitalised (Table 1).

View Table 1–2, Figure 1.

Three themes emerged from the analyses of clinical notes.

Theme 1: multiple factors contribute to delayed treatment

In 79% (27/34) of cases, clinicians considered leptospirosis as a potential cause of undifferentiated fever and included it in their initial diagnostic differential, often as the most likely diagnosis, and sometimes the only diagnosis.

“PC [presenting complaint]: headaches nausea and vomiting fevers, work exposure—deer urine splashed in his face over a week ago. Works in yards at venison processing plant… IMP [impression] leptospirosis.” (case_400)

However, diagnostic delays occurred when occupational or environmental exposure histories were not collected, or when clinicians failed to recognise the relevance of these exposures. Exposure history affecting clinical reasoning is demonstrated by case_360, where the primary physician did not record exposure history, and the patient was referred to the emergency department with cholangitis and sepsis as the differential diagnosis. The subsequent clinician obtained an exposure history and documented leptospirosis at the top of the problem list.

“Farmer high risk factors for leptospirosis.” (case_360)

In cases without the exposure history, a diagnosis was eventually made, often as part of a hepatitis screen, resulting in delayed treatment. Records of patient- or family-initiated diagnostic were rare, and were suggested by three cases.

“Patient concerned might have leptospirosis.” (case_303)

The presentation of severe symptoms, combined with the reliance on narrow textbook definitions, contributed to diagnostic uncertainty and delayed treatment. Leptospirosis was seldom described as a flu-like illness and often presented with severe symptoms consistent with sepsis. This resulted in the initiation of established sepsis protocols.

“Sepsis? Source—possible leptospirosis vs other.” (case_333)

“Blood cultures added, and antibiotics charted as he meets sepsis criteria.” (case_300)

In addition, the intensity of headaches prompted some clinicians to initially investigate for intracranial lesions using computed tomography (CT) scans before considering leptospirosis.

“He was investigated for meningitis with CT and LP [lumbar puncture] and blood work ...  His wife has asked if we tested for leptospirosis.” (case_408)

Clinicians were also misled when textbook features of leptospirosis—such as jaundice, conjunctival suffusion or thrombocytopenia—were absent.

“But patient has never had … any eye-related (jaundice or conjunctival effusion).” (case_305)

In one case, leptospirosis was explicitly ruled out based on the absence of these signs, even when other clinical and exposure indicators were present.

“Unexpected results on bloods: elevated LFTs [liver function tests] showing both cytolysis and congestion, but normal WBC [white blood cells] so it is more likely to be viral than bacterial? High count (usually for lepto it is a thrombocytopenia).” (case_393)

Theme 2: challenges with diagnosis and use of medications

Clinical notes revealed acute laboratory biochemical and haematological patterns. These included: renal impairment; hepatic impairment of obstructive, hepatocellular and mixed type without jaundice; normalising C-reactive protein and improving renal function; sterile haematuria in mid-stream urine samples; and subdued neutrophils and low lymphocytes in the clinical setting of a patient being investigated for bacterial sepsis. This study also found that mixed liver enzyme levels remained elevated, and continued to rise, even as clinical symptoms improved.

Most patients were treated on clinical suspicion with oral doxycycline and rehydration. In 35% (12/34) of cases, intravenous (IV) antibiotics (primarily cephalosporins) were administered. Two of these were for clinically suspected leptospirosis, the rest given as empiric stat dose in accordance with sepsis protocols. IV antibiotics were not used for more than 24 hours. One patient showed a clear clinical benefit; three others had increased rigors and chills following IV cephalosporin administration, suggestive of a possible Jarisch–Herxheimer reaction; however, this was not documented in any notes.

“Pt reviewed (following iv “empiric ceftriaxone”) 1445: now rigoring, T38.5 [temperature], feeling awful but no other new symptoms.” (case_333)

Comments on the effect of oral antibiotics varied: some described improvement by the following day while others indicated a slower response, and many notes lacked follow-up detail.

“Wife left a message concerned ABs aren’t working fast enough.” (case_356)

“Feels better still a bit of headache.” (case_409)

In 26% (9/34) of cases with available management information, clinical improvement was clearly linked to IV fluids alone. One patient continued to deteriorate despite antibiotic treatment, which was thought to be driven by dehydration and required intensive care unit (ICU) admission.

“ICU impression was that the episode was driven primarily by severe dehydration, his antibiotics were stopped at this point.” (case_377)

Treatment complexities arose when antibiotics were prescribed up to 6 weeks after symptom onset, upon receipt of definitive test results and after the acute phase illness had resolved. Clinicians offered antibiotics despite full resolution of symptoms.

“But [patient] would feel more comfortable covering with antibiotic.” (case_370)

The lack of clarity of the post-acute antibiotic role was also noted from public health services.

[Public Health offered antibiotics 6 weeks after onset of symptoms] “Patient declined as felt better.” (case_342)

Theme 3: inconsistent use and interpretation of confirmatory tests

This theme presented the most complexity, with unexpected references related to test requests within the text data. Of the 33 records with confirmatory testing information, approaches to the number and types of tests requested varied widely (Figure 1).

The timing between acute and convalescent sampling was not well understood. For example, a convalescent serology was requested 2 days after the acute sample, and not all convalescent serology requests were performed, leading to no serovar identification.

“Patient did not get blood test for second serology done.” (case_400)

“Spoke with Public Health and they will follow up with [patient] and look into whether strain identification is required.” (case_302)

In addition, clinicians received inconsistent advice from laboratories on the timing and interpretation of testing. These examples offer slightly different guidance on the timing of convalescent serology testing.

“Please send follow-up sample at least 10 days after first sample for further testing.” (case_397)

“Collect samples <7 days and >14 days from onset and further samples up to 90 days.” (case_405)

Further examples offered different recommendation on the timing of PCR tests.

“PCR on blood and urine are highly specific within the first 2 weeks of illness.” (case_353)

“Within 5 days of onset of symptoms blood.” (case_300)

“In the ‘immune phase’ 7–10 days from onset, urine is the sample of choice for PCR.” (case_300)

Test interpretation also affected decisions on work clearance and compensation, with many clinicians uncertain about infectiousness or serovar relevance.

“Also does he need to be clear of the infection before going back to work to prevent transmission? I don’t know much about this.” (case_350)

“Wondering if he should go to have blood test now, work wants to know the strain of lepto.” (case_356)

Employers and public health authorities sometimes imposed restrictions or gave unclear/inaccurate advice.

“…not allowed in dairy sheds when has lepto.” (case_340)

“Public Health phd [phoned] stating Pt [patient] needs a blood test around the 1/9/20 to make sure he has cleared lepto infection.” (case_353)

For case_353, the serology for Hardjo declined from 1:3,200 to 1:1,600 over a month following the acute presentation. This prompted the public health nurse to contact the medical centre, questioning if the results indicated a new infection or relapse. Miscommunication around public health implications extended to hospital specialists.

“The microbiologist requested convalescent sample to type and inform MPI [Ministry for Primary Industries].” (case_304)

“I rang Public Health during consult, and they already have his name re:lepto and they will contact WorkSafe for? ACC [Accident Compensation Corporation] and also will ring him as well.” (case_409)

Discussion

This study analysed clinical notes of diagnosed leptospirosis cases to understand the complexities surrounding clinical diagnosis and management. Results indicate that local disease presentation often diverged from global descriptions. Cases in this study were predominantly moderate to severely ill, even among those not requiring hospitalisation, with approximately 60% requiring hospitalisation. Few cases had their illness described as flu-like, and jaundice and conjunctival suffusion were also uncommon. This contrasts with previous studies suggesting that up to 90% of cases are mild and flu-like,^16,17 which may reflect that milder, self-limiting cases are less likely to seek medical attention. Differences in clinical presentation may also reflect variation in circulating serovars. In New Zealand, serovars from Leptospira interrogans are associated with more severe disease and higher hospitalisation rates compared with Leptospira borgpetersenii.¹⁸ Similarly, an Australian study attributed the absence of jaundice to the lack of serovar Icterohaemorrhagiae,¹⁹ which is also not present in New Zealand. Influenza was included in the differential diagnosis in only three cases, and in one case the clinician was misled due to concurrent illness in the patient’s family. Notably, the severity of headaches also separated leptospirosis from influenza.^10,12

Clinicians familiar with leptospirosis epidemiology suspected leptospirosis early on, based on key risk factors around occupational exposure and rural settings,¹³ compared to those who omitted exposure history. This supports prior studies on the importance of exposure assessment.^6,10,20 However, patients presenting outside traditionally recognised high-risk groups (e.g., urban population), or those exposed to newly identified risk factors in New Zealand such as flooding,²¹ may be at risk of under-diagnosis if these factors are not recognised by clinicians. Guidelines should be revised to provide clinicians with up-to-date risk factors.

Haake and Levett describe a septic process for leptospirosis and this study supports that, with most patients presenting signs resembling sepsis. While conjunctival suffusion and jaundice were uncommon, consistent with previous studies in New Zealand,^9,12 some clinicians sought textbook presentations for leptospirosis. Rigid criteria may create diagnostic distractions; thus, guidelines should avoid describing leptospirosis as influenza-like, reduce the emphasis on jaundice and conjunctival suffusion and highlight leptospirosis as an undifferentiated febrile illness ranging from mild severity to sepsis. Earl et al.’s recommendation to perform PCR, and acute and convalescent serology on all at-risk patients with a flu-like illness,⁹ should be extended to those presenting with sepsis.

Acute investigations more readily available in rural or resource-poor settings may have prognostic uses to aid in early diagnosis of leptospirosis, which is otherwise a clinical diagnosis.^9,22 While renal and hepatic impairment were commonly used indicators, others were not. This study reinforces previous findings linking leptospirosis with hyponatremia, lymphopenia, thrombocytopenia and anaemia.^2,10,11 De Silva noted that “neutrophil leukocytosis was not a prominent feature” of leptospirosis.²³ Our findings support a subdued neutrophil acute response to a septic process as an indicator of leptospirosis. This atypical haematological profile may help distinguish it from other bacterial causes of sepsis. A novel finding was transient liver function deterioration despite clinical improvement. These indicators may have prognostic value that warrants further comparative studies.^24,25

Once diagnosed, acute treatment followed national guidelines of oral doxycycline and rehydration,¹⁶ with few patients receiving IV antibiotics after diagnosis. Suputtamongkol et al. highlights the lack of consensus regarding optimal treatment for severe leptospirosis,²⁶ while Haake and Levett acknowledge limitations in existing antibiotic studies and the reduced likelihood of future placebo-controlled trials.² While this study was not designed to evaluate if clinical improvement was directly due to antibiotics or IV hydration, patient improvement following IV hydration was observed, aligning with WHO advice of “aggressive supportive care”;³ however, evidence regarding optimal fluid management in severe infection remains uncertain.²⁷

Complexities for treatment arose with post-acute diagnosis or prolonged symptoms with clinicians often deferring to patient preference for prescribing antibiotics. In one case, weeks post-onset, the patient reported subjective improvement with oral antibiotics. Doxycycline’s potential immunomodulatory effects may explain this response;²⁸ however, more research is needed. Current New Zealand guidelines cite WHO advice to offer antibiotics regardless of symptom duration, which in this study was up to 6 weeks and included asymptomatic patients.³ The basis of this recommendation requires further review.

Clinicians showed confidence diagnosing leptospirosis but inconsistent knowledge around diagnostic testing. This mirrors Irwin et al.’s findings that serological testing was often omitted or limited to acute-phase samples, leading to missed cases.¹⁰ Inconsistent communication from laboratories to clinicians further complicated diagnostic decision making, with variability in guidance and result interpretation likely contributing to underdiagnosis. Uncertainty around definitive testing and clarity on test timing relative to disease progression, appropriate test selection, result interpretation and the broader implications of test outcomes affected clinical care as well as long-term management, including decisions around compensation and work clearance. Clearer guidance and alignment of diagnostic criteria are needed to support both clinical and occupational health decision making. Standardising laboratory reporting and messaging nationwide is recommended, both to improve consistency and to support appropriate test use and follow-up.

Case follow-up notes revealed limited understanding among clinicians and public health advisers regarding their roles and responsibilities in the notification of leptospirosis to other authorities. Notably, some were unaware that patient consent was required before notifying WorkSafe New Zealand, and that the Ministry for Primary Industries is only to be informed of exotic serovars. These regulatory requirements underscore the importance of accurate diagnostic testing and the need to protect patient confidentiality to avoid deterring healthcare engagement. Providing guidance around regulatory obligations for managing human leptospirosis in New Zealand is recommended.

Study limitations

This study is limited by varied and incomplete clinical notes, which lacked serial, standardised clinical enquiry and investigations, making comparisons between patients difficult and limiting the ability to draw conclusions about antibiotic treatment outcomes.

This study is also limited to patients who sought medical care, were suspected of leptospirosis and had a positive test, introducing selection bias. As such, the findings may not be generalisable to undiagnosed cases, including those who were not tested, were treated empirically or experienced mild illness and did not seek care.

Previous studies suggest infecting serovar can influence clinical presentation;^9,12,18–20 therefore, the findings may be most applicable to the New Zealand context or regions with similar serovars.

Conclusion

This paper provides insights into the clinical presentation, diagnostic challenges and management of human leptospirosis in New Zealand. While clinicians demonstrated confidence in considering leptospirosis when faced with an undifferentiated fever case who had a positive exposure history, complexities arose with definitive testing, result interpretation and post-acute management. Additionally, knowledge gaps were noted regarding use of laboratory parameters in the acute setting as well as the possibility that textbook descriptions of leptospirosis could be misleading. Based on the findings, this study makes several recommendations to improve early diagnosis and management of leptospirosis while addressing gaps in clinician knowledge and public health response (Table 2).