Advertisement
  • Loading metrics

Risk factors for human acute leptospirosis in northern Tanzania

  • Michael J. Maze ,

    Roles Formal analysis, Investigation, Methodology, Writing – original draft

    Michael.Maze@cdhb.health.nz

    Affiliations Centre for International Health, University of Otago, Dunedin, New Zealand, Kilimanjaro Christian Medical Centre, Moshi, Tanzania

  • Shama Cash-Goldwasser,

    Roles Formal analysis, Methodology, Writing – review & editing

    Affiliations Kilimanjaro Christian Medical Centre, Moshi, Tanzania, Duke Global Health Institute, Duke University, Durham, North Carolina, United States of America

  • Matthew P. Rubach,

    Roles Formal analysis, Investigation, Writing – review & editing

    Affiliations Kilimanjaro Christian Medical Centre, Moshi, Tanzania, Duke Global Health Institute, Duke University, Durham, North Carolina, United States of America, Division of Infectious Diseases, Duke University Medical Center, Durham, North Carolina, United States of America

  • Holly M. Biggs,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliation Division of Infectious Diseases, Duke University Medical Center, Durham, North Carolina, United States of America

  • Renee L. Galloway,

    Roles Investigation, Methodology, Resources, Writing – review & editing

    Affiliation Bacterial Special Pathogens Branch, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America

  • Katrina J. Sharples,

    Roles Formal analysis, Methodology, Supervision, Writing – review & editing

    Affiliations Centre for International Health, University of Otago, Dunedin, New Zealand, Department of Mathematics and Statistics, University of Otago, Dunedin, New Zealand

  • Kathryn J. Allan,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliation Boyd Orr Centre for Population and Ecosystem Health, Institute of Biodiversity, Animal Health and Comparative Medicine, University of Glasgow, Glasgow, United Kingdom

  • Jo E. B. Halliday,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliation Boyd Orr Centre for Population and Ecosystem Health, Institute of Biodiversity, Animal Health and Comparative Medicine, University of Glasgow, Glasgow, United Kingdom

  • Sarah Cleaveland,

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Supervision, Writing – review & editing

    Affiliation Boyd Orr Centre for Population and Ecosystem Health, Institute of Biodiversity, Animal Health and Comparative Medicine, University of Glasgow, Glasgow, United Kingdom

  • Michael C. Shand,

    Roles Data curation, Methodology, Software, Supervision, Writing – review & editing

    Affiliation Boyd Orr Centre for Population and Ecosystem Health, Institute of Biodiversity, Animal Health and Comparative Medicine, University of Glasgow, Glasgow, United Kingdom

  • Charles Muiruri,

    Roles Methodology, Project administration, Writing – review & editing

    Affiliation Division of Infectious Diseases, Duke University Medical Center, Durham, North Carolina, United States of America

  • Rudovick R. Kazwala,

    Roles Funding acquisition, Investigation, Writing – review & editing

    Affiliation Department of Veterinary Medicine and Public Health, Sokoine University of Agriculture, Morogoro, Tanzania

  • Wilbrod Saganda,

    Roles Funding acquisition, Investigation, Writing – review & editing

    Affiliation Mawenzi Regional Referral Hospital, Moshi, Tanzania

  • Bingileki F. Lwezaula,

    Roles Investigation, Writing – review & editing

    Affiliation Mawenzi Regional Referral Hospital, Moshi, Tanzania

  • Blandina T. Mmbaga,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliations Kilimanjaro Christian Medical Centre, Moshi, Tanzania, Division of Infectious Diseases, Duke University Medical Center, Durham, North Carolina, United States of America, Kilimanjaro Christian Medical University College, Moshi, Tanzania, Kilimanjaro Clinical Research Institute, Moshi, Tanzania

  • Venance P. Maro,

    Roles Conceptualization, Funding acquisition, Investigation, Writing – review & editing

    Affiliations Kilimanjaro Christian Medical Centre, Moshi, Tanzania, Kilimanjaro Christian Medical University College, Moshi, Tanzania

  •  [ ... ],
  • John A. Crump

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – review & editing

    Affiliations Centre for International Health, University of Otago, Dunedin, New Zealand, Kilimanjaro Christian Medical Centre, Moshi, Tanzania, Duke Global Health Institute, Duke University, Durham, North Carolina, United States of America, Division of Infectious Diseases, Duke University Medical Center, Durham, North Carolina, United States of America, Kilimanjaro Christian Medical University College, Moshi, Tanzania

  • [ view all ]
  • [ view less ]

Risk factors for human acute leptospirosis in northern Tanzania

  • Michael J. Maze, 
  • Shama Cash-Goldwasser, 
  • Matthew P. Rubach, 
  • Holly M. Biggs, 
  • Renee L. Galloway, 
  • Katrina J. Sharples, 
  • Kathryn J. Allan, 
  • Jo E. B. Halliday, 
  • Sarah Cleaveland, 
  • Michael C. Shand
PLOS
x

Abstract

Introduction

Leptospirosis is a major cause of febrile illness in Africa but little is known about risk factors for human infection. We conducted a cross-sectional study to investigate risk factors for acute leptospirosis and Leptospira seropositivity among patients with fever attending referral hospitals in northern Tanzania.

Methods

We enrolled patients with fever from two referral hospitals in Moshi, Tanzania, 2012–2014, and performed Leptospira microscopic agglutination testing on acute and convalescent serum. Cases of acute leptospirosis were participants with a four-fold rise in antibody titers, or a single reciprocal titer ≥800. Seropositive participants required a single titer ≥100, and controls had titers <100 in both acute and convalescent samples. We administered a questionnaire to assess risk behaviors over the preceding 30 days. We created cumulative scales of exposure to livestock urine, rodents, and surface water, and calculated odds ratios (OR) for individual behaviors and for cumulative exposure variables.

Results

We identified 24 acute cases, 252 seropositive participants, and 592 controls. Rice farming (OR 14.6), cleaning cattle waste (OR 4.3), feeding cattle (OR 3.9), farm work (OR 3.3), and an increasing cattle urine exposure score (OR 1.2 per point) were associated with acute leptospirosis.

Conclusions

In our population, exposure to cattle and rice farming were risk factors for acute leptospirosis. Although further data is needed, these results suggest that cattle may be an important source of human leptospirosis. Further investigation is needed to explore the potential for control of livestock Leptospira infection to reduce human disease.

Author summary

Leptospirosis is an under-recognized but important cause of febrile illness and death in Africa. The bacteria that cause leptospirosis have their usual life cycle in animals; humans are infected as accidental hosts. There is considerable variation between countries as to which reservoir animals and human activities are important for transmission of leptospirosis to humans. In many tropical countries flooding and rodents are the dominant sources of human infection. However, in Africa it is unknown which sources of leptospirosis are most responsible for human infection and what behaviors put people at risk for infection We performed a prospective cross-sectional study, to identify risk factors for acute leptospirosis and sources of human infection. We identified contact with cattle and work in rice fields as risk factors for acute leptospirosis. Our findings indicate that cattle may be an important source for human leptospirosis, and therefore control of leptospirosis in livestock may help prevent leptospirosis in people. Further work is needed to isolate Leptospira from humans and livestock. Rice farming was an uncommon activity in our study, but strongly associated with acute leptospirosis. Efforts are warranted to prevent infection in rice farmers living in Africa.

Introduction

Leptospirosis is a zoonotic bacterial infection and is increasingly recognized as an important cause of fever in Africa [1]. Leptospirosis was a leading cause of severe febrile illness in a study conducted in northern Tanzania during 2007–8, where it was diagnosed in 8.8% of participants [2]. The annual incidence of severe acute leptospirosis in northern Tanzania is high, but has fluctuated during surveillance over two time periods: from 75–102 cases per 100,000 people in 2007–08 to 11–18 cases per 100,000 people in 2012–14, suggesting dynamic transmission patterns [3]. An understanding of major animal reservoirs, sources, and modes of transmission to humans is required to inform leptospirosis control.

Animals infected by Leptospira may become carriers and excrete Leptospira in urine leading to environmental contamination. Humans can be infected following direct exposure to the urine of infected animals or through contact with contaminated surface water or moist soil [5]. Portals of entry include mucous membranes and broken skin [5]. While the major reservoirs, sources of human infection, and modes of transmission of infection are established on a global scale, there is substantial variation by location reflecting the diverse ecology of Leptospira. In many tropical countries, rodent species are considered the most important animal reservoir for human infection [4]. As such, dominant risk factors for leptospirosis in many tropical countries include activities that expose individuals to rodent urine, such as living in urban slums, proximity to sewers, and exposure to flood waters [4, 6, 7]. In Tanzania and most other African countries, the risks factors for human infection are not well characterized [1, 4], and there is some evidence that the risk factors may differ from other tropical countries. In northern Tanzania there is evidence that leptospirosis is more common in rural areas where both livestock and rodents could be important sources of human infection [8], and previous Leptospira exposure studies have identified livestock farmers as a high-risk group for Leptospira seropositivity [9]. Serogroup reactivity patterns of acute human leptospirosis infections have also suggested that livestock may be reservoirs for human cases [8], and studies of livestock have found high proportions that were seropositive or with leptospiruria [1012]. To inform leptospirosis control in Tanzania, we aimed to identify risk factors for acute leptospirosis and Leptospira seropositivity, and identify sources of human Leptospira infection.

Methods

Study setting

We conducted a cross-sectional study at Kilimanjaro Christian Medical Centre (KCMC), a 450-bed zonal referral hospital and, Mawenzi Regional Referral Hospital (MRRH) a 300-bed regional referral hospital, both in Moshi. Moshi (population ~180,000) is the administrative capital of the Kilimanjaro Region (population ~1.6 million) of Tanzania. Moshi is situated at approximately 890 meters above sea level and has a tropical climate with rainy seasons from October through December, and March through May. Agriculture in northern Tanzania includes smallholder systems involving mixed crop and livestock farming, as well as pastoralism.

Study procedures and participants

We enrolled pediatric and adult patients presenting to KCMC and MRRH from February 2012 through May 2014. From Monday through Friday, we screened all patients in the adult medical ward at KCMC and the adult and pediatric medical wards at MRRH within 24 hours of admission, as well as patients presenting to the outpatient department at MRRH. We enrolled consecutive eligible inpatients and every second eligible outpatient. Patients were eligible to participate if they had an axillary temperature of >37.5°C or a tympanic, oral, or rectal temperature of ≥38.0°C at presentation. Inpatients were also eligible if they reported a history of fever within the past 72 hours. After obtaining informed consent, a trained study team member completed standardized clinical history and risk factor questionnaires. The risk factor questionnaire included questions on socio-demographic characteristics, participant living environment, and daily activities performed over the past 30 days, focusing specifically on animal-related activities, exposure to surface water and to rodents (S1 Text). The questionnaire was designed to include established risk factors for leptospirosis from studies done in other settings [4, 6, 7, 1315], and was piloted prior to use. For participants who lived in the Kilimanjaro Region, study personnel visited participant households to record Global Positioning System (GPS) coordinates of participants’ dwellings. Clinician diagnoses were recorded. Participants were asked to return 4–6 weeks after enrollment for collection of a convalescent serum sample.

Laboratory methods

Blood was allowed to clot for between 30 and 60 minutes. It was then centrifuged for 15 minutes at 1,126–1455 relative centrifugal force to separate serum. Serum was stored at -80°C. Serum specimens were batch shipped on dry ice from Moshi, Tanzania to Atlanta, GA, United States of America for testing. Serology for leptospirosis was performed at the US Centers for Disease Control and Prevention using the standard microscopic agglutination test (MAT) with a panel of 20 Leptospira serovars belonging to 17 serogroups [16]. These included: Australis (represented by L. interrogans serovar Australis, L. interrogans serovar Bratislava), Autumnalis (L. interrogans serovar Autumnalis), Ballum (L. borgpetersenii serovar Ballum), Bataviae (L. interrogans serovar Bataviae), Canicola (L. interrogans serovar Canicola), Celledoni (L. weilii serovar Celledoni), Cynopteri (L. kirschneri serovar Cynopteri), Djasiman (L. interrogans serovar Djasiman), Grippotyphosa (L. interrogans serovar Grippotyphosa), Hebdomadis (L. santarosai serovar Borincana), Icterohaemorrhagiae (L. interrogans serovar Mankarso, L. interrogans Icterohaemorrhagiae), Javanica (L. borgpetersenii serovar Javanica), Mini (L. santarosai serovar Georgia), Pomona (L. interrogans serovar Pomona), Pyrogenes (L. interrogans serovar Pyrogenes, L. santarosai serovar Alexi), Sejroe (L. interrogans serovar Wolffi), and Tarassovi (L. borgpetersenii serovar Tarassovi). MAT was performed beginning at a dilution of 1:100, with subsequent two-fold dilutions. Positive and negative controls were included with each run.

Case definitions

We defined leptospirosis cases as participants with either a four-fold rise in agglutinating antibody titers between acute and convalescent serum, or a single reciprocal titer of ≥800 [17]. Seropositivity was defined as a single positive reciprocal titer of ≥100 from either sample. Controls were participants with negative titers on both acute and convalescent serum samples. The predominant reactive serogroup for cases and seropositive participants was defined as the serogroup containing the serovar with the highest titer.

Geospatial and rainfall data

For each participant, village population density was calculated from the 2012 Tanzania Population and Housing Census [18]. For the purpose of analysis, a priori zone classifications were applied to each village [19]. Villages with a population density of 10 inhabitants/km2 were classified as urban; villages ≤15km distance from urban areas with a population density ≥3 and < 10 inhabitants/km2 were classified as peri-urban; and villages ≥15km distance from an urban area with a population density of <3 inhabitants/km2 [19]. Georeferenced mean annual rainfall and soil type data were obtained from the 2002 Kenya International Livestock Research Institute report [20]. Land use data were obtained from the 2010 National Geomatics Center of China report [21]. Daily rainfall data were obtained from the Tanzania Production Company (TPC) rainfall stations located near Moshi.

Statistical analysis

Patient history, questionnaire, and MAT data were entered using the Cardiff Teleform system (Cardiff, Inc., Vista, CA, USA) into an Access database (Microsoft Corporation, Redmond, WA, USA). Geospatial data were managed using QGIS, version 2.8.3 (Free Software Foundation, Boston, MA, USA). Spatial scan statistics were calculated using a Bernoulli model to assess evidence of spatial clustering of cases using SatScan version 9.0 (www.satscan.org) [22]. All other analyses were performed using Stata, version 13.1 (StataCorp, College Station, TX, USA).

Modeling strategy.

Logistic regression was used to investigate associations between independent variables and two separate outcome measures: acute leptospirosis and Leptospira seropositivity. Initially all associations between individual behavior variables and our outcome variables were assessed by bivariable logistic regression. In addition, to understand the relationship between independent variables, we performed bivariable logistic regression between all independent variables. We then developed models to investigate the behavioral variables and the geospatial variables separately. Because of the high ratio of independent variables to cases of acute leptospirosis, we considered multivariable logistic regression models of the individual behavior variables, and the georeferenced variables to be unstable. Therefore, to facilitate multivariable modeling, and to assess the effects of cumulative exposure we created summary scales to quantitatively estimate overall patient exposure to each of the main modes and sources of infection that we identified from the published literature: urine of cattle, goats, pigs and rodents, and surface water [4]. This meant that when analyzing risk factors for acute leptospirosis we built a single multivariable model using the exposure scores that are described in detail below. When analyzing risk factors for Leptospira seropositivity, we were able to build three multivariable models: one multivariable model using individual behavioral risk factors, one model using individual geospatial risk factors, and one model using the exposure scores.

Development of exposure scores.

We used an analytic hierarchy process to develop these scales [23]. First we identified relevant behaviors and living conditions from the risk factor questionnaire to be included in each scale. We then identified locally experienced subject matter experts, including livestock field officers, physicians, rodent ecologists, veterinarians, water engineers, water and sanitation epidemiologists, and zoonotic disease epidemiologists. For each potential source of infection we asked experts to rank each relevant behavior against every other behavior in terms of the likelihood and intensity of exposure to the source using a 9-point bidirectional scale. We then calculated weightings using a matrix (Appendices 2–4) that added, for each behavior, reciprocals of the score from each pairwise comparison [24]. We assessed precision and error in the judgment process through measuring an expert’s internal consistency in multiple pairwise comparisons. We excluded weightings by experts who provided internally inconsistent answers, as designated by a consistency ratio >0.2 [24], and then calculated the geometric mean of the weightings given by all included experts. To aid interpretation of exposure scores, we multiplied all weights within each scale by a constant, so that possible scores on each scale ranged from 0 to 5. Finally, we derived an overall score for each participant, on each exposure scale, based on their questionnaire answers, such that someone who had performed none of the exposure activities scored zero, and someone who performed all of the activities scored 5.

Model building using exposure scales.

Relationships between exposure scores and log odds of acute leptospirosis and Leptospira seropositivity, were assessed using fractional polynomial transformations of the exposure scales [25]. We allowed up to two degrees of freedom, and utilized the function selection procedure of Stata’s multivariable fractional polynomial algorithm to select the best fitting transformation for each exposure scale. The correlation between exposure scales was assessed using linear regression. Interactions between exposure scales were assessed using factor variables. Through use of directed acyclic graphs, we considered that all exposure scores might act as confounders, and our initial multivariable models included all exposure scales. Variables were examined for co-linearity using variance inflation factors. We used stepwise backwards elimination to arrive at the model that minimized the Akaike Information Criterion (AIC) [26], p values were two sided, and the significance level was set at 0.05.

Model building using individual behaviors, and geospatial risk factors.

Our initial models of individual behavioral and our initial models of geospatial risk factors included all variables with a p value <0.2 in bivariable logistic regression. No variables were forced into the model as confounders. We performed backwards stepwise model selection to minimize the AIC.

Research ethics

This study was conducted in accordance with the Declaration of Helsinki. It was approved by the KCMC Research Ethics Committee (#295), the Tanzania National Institute for Medical Research National Ethics Coordinating Committee (NIMR1HQ/R.8cNo1. 11/283), Duke University Medical Center Institutional Review Board (IRB#Pro00016134), and the University of Otago Human Ethics Committee (Health) (H15/055). Written informed consent was obtained from all participants or their guardians.

Results

Enrollment and participant characteristics

Of 15,305 patients admitted and 30,413 presenting to the outpatient department, 2,962 met eligibility criteria and 1,416 (47.8%) were enrolled. Of 1,293 participants who completed the risk factor questionnaire and had serum tested, 24 (1.9%) met the study criteria for acute leptospirosis, 252 (19.5%) were seropositive, and 592 (45.8%) were classified as controls (Fig 1). The remaining 449 (34.7%) were seronegative but provided only a single serum sample and so were excluded from analysis. The frequency with which participants were predominantly reactive to different serogroups is shown in Table 1. Participant characteristics are shown in Table 2. Clinicians did not diagnose leptospirosis in any study participant. Four (25.0%) of 16 leptospirosis cases with discharge diagnoses recorded were diagnosed with malaria despite negative blood parasite examinations.

thumbnail
Fig 1. Study flow diagram for patients seeking care at Kilimanjaro Christian Medical Centre and Mawenzi Regional referral hospital in Moshi, Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.g001

thumbnail
Table 1. Predominantly reactive serogroup of leptospirosis cases and participants seropositive to Leptospira, northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.t001

thumbnail
Table 2. Demographic and clinical characteristics of study participants, northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.t002

Association of risk factors

Bivariable logistic regression of individual risk factors are included in S2 Table. There was a strong association between behaviors involving a single livestock species. For example having cleaned cattle waste was associated with having fed cattle with an OR 324.1 (95% confidence intervals 96.6–1087.0). There was some association between behaviors involving different livestock species. For example having cleaned cattle waste was associated with having cleaned goat waste with an OR 28.8, 95% confidence interval 12.0–69.1. There was a small magnitude association between rodent contact variables and livestock related variables. For example owning cattle was not associated with seeing rodents frequently in the house, compound or fields, and had a low magnitude association with seeing rodents in the kitchen or food store (OR 1.5, 95 confidence intervals 1.1–2.1).

Acute leptospirosis individual behaviors

Results for the logistic regression analysis of individual behaviors are shown in Table 3. On bivariable regression, variables associated with acute leptospirosis included working in rice fields (OR 14.6, 95% confidence intervals (CI) 2.9–59.5); cleaning up cattle waste (OR 4.3, CI 1.2–12.9); feeding cattle (OR 3.9, CI 1.3–10.3) and working as a farmer (OR 3.3, CI 1.3–8.2).

thumbnail
Table 3. Bivariable logistic regression of individual risk factors for acute leptospirosis among patients with febrile illness (N = 616), northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.t003

Acute leptospirosis exposure scales

Nine (42.9%) of 21 experts (three livestock field officers, four veterinarians, and two zoonotic disease epidemiologists provided internally consistent multiple pairwise rankings of the relative exposure to livestock urine from the behaviors listed in Table 4. Four (100.0%) of four experts (one water engineer, one water and sanitation epidemiologist, and two zoonotic disease epidemiologists) provided consistent multiple pairwise rankings of the relative exposure to surface water. Three (75.0%) of four experts (one rodent ecologist, one veterinarian, and one zoonotic disease epidemiologist) provided consistent multiple pairwise rankings of the relative exposure to rodent urine. The individual behaviors evaluated for each exposure scale and the geometric means of the weights assigned to each are listed in Table 4. The results of pairwise comparisons, and calculated weights for each behavior are presented in S3 Table, S4 Table, and S5 Table. The distributions of participants’ exposure scores on each scale are shown in Fig 2. Overall, 534 (69.3%) of participants had no evidence of exposure to cattle urine, 563 (73.0%) had no exposure to goat urine, 241 (31.2%) had no exposure to rodent urine, and 262 (34.0%) had no exposure to surface water. There was limited correlation between cattle urine exposure and both goat urine exposure (r2 = 0.21) and pig urine exposure (r2 = 0.04). In addition there was little correlation between livestock urine exposure scores and rodent urine exposure (for example, cattle urine exposure and rodent urine exposure, r2 = 0.04), livestock exposure scores and surface water exposure (for example cattle urine and surface water (r2 = 0.02), and between rodent urine exposure and surface water exposure (r2 = 0.02). All exposure scales had a linear relationship with log odds of acute leptospirosis

thumbnail
Fig 2. Participant scores of exposure to animal urine and surface water, northern Tanzania, 2012–14 (N = 844).

https://doi.org/10.1371/journal.pntd.0006372.g002

thumbnail
Table 4. Component risk factors and relative weights for exposure to multiple leptospirosis infection sources derived from an analytic hierarchy process conducted among East African subject matter experts, 2015.

https://doi.org/10.1371/journal.pntd.0006372.t004

Our bivariable logistic regression (Table 5) found that increasing exposure to cattle urine (OR 2.3, CI 1.1–4.7) and exposure to rodents (OR 1.7, CI 1.1–2.8) were both associated with increased odds of acute leptospirosis. In multivariable logistic regression (Table 5), no exposure scale was independently associated with leptospirosis. As shown in S6 Table, there were no significant interactions. The largest variance inflation factor was 1.33.

thumbnail
Table 5. Bivariable and multivariable logistic regression models of association between exposure scales and acute leptospirosis among patients with febrile illness in northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.t005

Acute leptospirosis geospatial and temporal analysis

GPS co-ordinates were available for houses of 649 (84.2%) participants. No two or more participants lived at the same household. Land use designation could be determined from participant’s self-reported village of residence for an additional 79 (10.2%) participants. There was no evidence of clustering in the spatial distribution of cases. Results of the bivariable logistic regression analysis of geo-referenced variables and rainfall, and acute leptospirosis are shown in Table 6. There were no statistically significant associations.

thumbnail
Table 6. Bivariable logistic regression of temporal and geo-referenced risk factors for acute leptospirosis among patients with febrile illness, northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.t006

Leptospira seropositivity

Results of the logistic regression of individual risk factors for Leptospira seropositivity are listed in Table 7. Working in rice fields (OR 3.6, 95% CI 1.5–9.0); slaughtering goats (OR 2.3, 95% CI 1.0–4.8), working as a farmer (OR 1.8, 95% CI 1.3–2.5), and frequently seeing rodents in the kitchen (OR 1.5, 95% CI 1.1–2.1) were significant risk factors (p < 0.05) on bivariable regression. We fitted an initial multivariable model using the risk factors shown in Table 8. As shown in S6 Table, we did not identify any significant interactions between variables. In our final multivariable model, working as a farmer (OR 1.6, CI 1.1–2.3), working in the rice fields (OR 2.7 CI 1.0–7.2), or seeing rodents in the kitchen ≥ once per week (OR 1.5, CI 1.0–2.1) were all independent risk factors for Leptospira seropositivity. Walking barefoot (OR 0.7, CI 0.5–0.9) and owning dogs (OR 0.6, CI 0.4–1.0) were associated with reduced odds of Leptospira seropositivity.

thumbnail
Table 7. Bivariable logistic regression of risk factors for Leptospira seropositivity among patients with febrie illness in northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.t007

thumbnail
Table 8. Multivariable logistic regression of individual risk factors Leptospira seropositivity among patients with febrile illness in northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.t008

The logistic regression models of the exposure scales and Leptospira seropositivity are shown in Table 9. Increasing exposure to rodent urine (OR1.2, CI 1.0–1.5) was associated with Leptospira seropositivity on bivariable logistic regression, but not on multivariable regression.

thumbnail
Table 9. Bivariable and multivariable logistic regression models of association of exposure scales and Leptospira seropositivity among patients with febrile illness in northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.t009

Results of the bivariable logistic regression analysis of rainfall and Leptospira seropositivity are shown in Table 10. There was an inverse association with mean annual rainfall >1,600mm per year (OR 0.56, 95% CI 0.33–0.93). We fitted an initial multivariable model using household elevation, mean annual rainfall, maximum daily rainfall in the preceding 30 days, and total rainfall in the preceding 30 days. The final model contained elevation (OR 0.99 per 10m, CI 0.98–1.0, p = 0.06), and total rainfall in the preceding 30 days (OR 1.2 per 100mm, CI 0.95–1.5, p = 0.13) but neither association was statistically significant. An analysis of the risk factors for seropositivity against Leptospira serogroup Icterohaemorrhagiae is included as S6 Table.

thumbnail
Table 10. Bivaraite logistic regression of temporal and geo-referenced risk factors for Leptospira seropositivity among patients with febrile illness, in northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.t010

Discussion

We identified multiple associations between exposure to cattle and acute leptospirosis, suggesting that cattle may be important sources of human leptospirosis in northern Tanzania. We also identified work in rice fields as an important risk factor for human leptospirosis. These findings must be interpreted with caution, as they were based on a small number of cases, and were present in only bivariable regression. Despite this, our findings have implications for the control and prevention of leptospirosis in Tanzania.

On bivariable regression, exposure to cattle was associated with acute human leptospirosis both when we evaluated individual behaviors and scales of cumulative exposure to cattle urine. These findings support other data from northern Tanzania that indicate that livestock may be an important source of human leptospirosis [31]. Among cattle slaughtered for meat in the Moshi area, 7.6% of cattle tested were carrying pathogenic Leptospira spp. in their kidneys [31]. Furthermore, seroreactivity against serogroups Australis and Sejroe, the two dominant serogroups among human cases in our study, was also frequently observed among cattle slaughtered for meat in the Moshi area in 2014 [12]. Our findings are also consistent with studies examining risk factors for Leptospira seropositivity in Africa. Leptospira seropositivity was common among abattoir workers in Kenya and Tanzania [11, 27]. In rural Uganda, livestock skinning was reported as a risk factor for seroreactivity and human seropositivity to livestock-associated Leptospira serovars was common [28]. In a global context, cattle have also been identified as a key risk factor in other rural livestock-farming communities in Central America and South Asia [14, 15], suggesting that strategies to reduce either livestock leptospirosis or transmission of leptospirosis from livestock to humans may be important global public health interventions.

Rodent exposure is an important risk factor for leptospirosis in the tropics, particularly in urban areas of Asia and South America [4, 29, 30]. In our study, an increasing score on the exposure to rodent urine scale was associated with acute leptospirosis in bivariable regression. However, the only individual component of the scale for which we found an association on bivariable regression was smallholder farming. Since smallholder farming may involve substantial exposure to both livestock and rodents, and other rodent related variables were not associated with leptospirosis the role of rodents in this association is uncertain. We also found that frequently sighting rodents in the kitchen or food store was associated with Leptospira seropositivity. Rodents could transmit leptospirosis to humans, or act as a reservoir that transmit Leptospira to livestock. However, recent work in the Kilimanjaro Region found no evidence of Leptospira urinary shedding, or renal infection among 393 wild rodents [31]

Although practiced by few participants, we found an association between working in rice fields, and both acute leptospirosis and Leptospira seropositivity. In some areas of northern Tanzania rice farming is practiced intensively, and there are active efforts to increase irrigated, continuously flooded rice farming across Tanzania [32]. In Asia rice farming is an established risk factor for leptospirosis. In Asia humans are infected through prolonged contact with water that may be contaminated by infected animal hosts [4, 29]. Further work is needed to evaluate possible sources of contamination of rice paddies in Tanzania and promote personal protective measures among rice farmers.

We did not find associations between acute leptospirosis and rainfall, or environmental risk factors around the home. The small number of cases available for analysis, and the relative lack of resolution of geo-referenced data meant that this result must be interpreted with caution. The lack of association with heavy rainfall differs from findings of studies from other locations [33, 34]. We found that seropositivity was associated with lower elevation and lower rainfall. While we did not have household level slope data, the topography of the study area includes steeply sloping terrain on the flanks of Mount Kilimanjaro that may not favor surface water accumulation. The lack of association between leptospirosis and home location may indicate that the workplace is an important site for infection [9, 11]. Future studies should collect data regarding workplace location.

Clinicians did not diagnose leptospirosis during the study period, and over-diagnosis of malaria was common. At the time of our study, there were no locally available, accurate diagnostic tests for leptospirosis. In addition, despite the high incidence in the region, clinician awareness of leptospirosis and other zoonotic diseases remains low [35]. This highlights the need for clinician education and evaluations in Africa of inexpensive point-of-care diagnostic tests.

We found that risk factors and the pattern of predominant reactive serogroups among leptospirosis cases was markedly different from those in seropositive individuals, for whom the febrile illness concurrent with enrollment was unlikely to be leptospirosis. In particular, reactivity to serogroup Icterohaemorrhagiae was common among seropositive participants, but there were few acute cases associated with this serogroup. These results may indicate that a serovar from the Icterohaemorrhagiae serogroup was circulating in this region [36], causing only mild disease not requiring tertiary medical care. Elsewhere, a difference in severity of disease has been linked to variability of infecting Leptospira species [37], Alternatively, the presence of Icterohaemorrhagiae seropositivity but absence of acute cases could indicate historic circulation of this serogroup that has since declined. Other results suggest that leptospirosis has a dynamic epidemiology in this area with the emergence and decline of specific serovars over time [3]. Cross reactivity between serogroups, and non-specific reactivity are other possible explanations [38].

Our study had several limitations. First, the prevalence of acute leptospirosis was lower than anticipated [8], potentially curtailing our ability to detect important associations. Conversely, associations of individual activities and leptospirosis identified by this study were sometimes based on only a few cases and should be interpreted with caution, especially given the multiple statistical tests. In addition, changes in leptospirosis incidence in the study area might also reflect changes in predominant sources and modes of transmission over time [3]. Second, the associations for acute leptospirosis were seen only on bivariable analysis, and these associations may be due to confounding from unobserved behaviors. Due to the complex interconnection between individual behaviours, we also consider that confounding may influence the multivariable logistic regression model of individual behaviours and Leptospira seropositivity. For example, the inverse association of walking barefoot and leptospirosis is puzzling, and we think it is likely to be influenced by an association with some protective factor, despite not identifying such an association among the behaviors we investigated. Diagnostic test limitations may have also introduced classification errors of participant cases or controls into our analysis. Leptospirosis is notoriously difficult to diagnose, particularly in the acute stages of illness and all currently available diagnostic tests for leptospirosis, including MAT [39], are imperfect. The sensitivity of MAT on paired serum samples is approximately 80% and the specificity close to 100% [40]. Specifically, not all participants with leptospirosis will seroconvert [40], and it is not possible to differentiate between historic and recent infection based on a single high titer [41]. We chose MAT for our case definitions since MAT on paired serum samples, while imperfect, remains the reference standard [40]. Furthermore, culture, nucleic acid amplification and point-of-care IgM serology lack sensitivity in our setting [12, 42, 43], and reports from other settings have been mixed [39, 4446]. Our MAT panel comprising 20 serovars covered the major Leptospira serogroups that cause human disease, and all those within which African isolates are grouped [1]. We did not use locally isolated serovars and this may have influenced identification of cases. However, studies on the use of local isolates in MAT reference panels have shown that they do not necessarily perform better than other serovars from the same serogroup [47, 48]. Our analysis of acute leptospirosis was limited to cases across all serogroups. We acknowledge that risk factors may vary by infecting serovar, and pan-serogroup analyses may mask important associations.

We developed scales for use in our analyses for dimension reduction due to the unanticipated low number of cases. We suggest that cumulative exposure scales may have a future role in assessing sources of acute leptospirosis, as they allow assessment of cumulative exposure that may be important in assessing individual risk of disease. The analytic hierarchy process was an appropriate method of creating these scales, as it is an effective tool for quantifying multi-dimensional qualitative knowledge [24]. While we acknowledge that there is scope to improve our cumulative exposure scales, our scales that quantify expert opinion offer more biologically plausible groupings than statistical methods of dimension reduction. Key areas for future development of cumulative exposure scales are to validate them across multiple groups of experts, and to formally compare their effectiveness against purely statistical dimension reduction. Since our questionnaire sought exposures over a 30 day period, recall bias may have influenced our findings. Finally, we enrolled only 47.1% of eligible patients. We found no bias towards particular ethnic or occupational groups. However, we cannot rule out the possibility that the enrollment pattern influenced our results. Despite these limitations, the consistency of the association of the livestock related variables strengthens our confidence in the interpretation of their role in transmitting leptospirosis to people in our region.

Our results have implications for control of leptospirosis. Transmission of leptospirosis within rice fields, and from livestock to people is amenable to control through personal protective equipment for those performing high risk activities [49]. In addition, Leptospira vaccines are available for use in livestock against some Leptospira serovars. In some countries such vaccines have contributed to successful control of leptospirosis [49]. However, before a vaccination program is considered it is essential to understand reservoir structure and predominant infecting serovars.

Our study identifies associations between cattle contact and work in rice fields with acute leptospirosis. Our findings suggest that cattle may be a source of human leptospirosis in northern Tanzania. Further work is needed to determine if these findings are stable over time, and to investigate the link by isolating infecting serovars from humans and animal hosts. The development of local MAT capacity, or use of nucleic acid amplification or point-of-care IgM tests that have sufficiently high sensitivity would enable real-time diagnosis and allow testing of potential animal hosts living in proximity to humans with acute leptospirosis. Nonetheless, our findings suggest that control of Leptospira infection in livestock could play a role in preventing human leptospirosis in Africa.

Supporting information

S1 Text. Participant questionnaire assessing risk factors for zoonotic disease among patients with febrile illness in northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.s001

(PDF)

S1 Table. Spearman’s correlation coefficients for individual behavior variables among patients with febrile illness in northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.s002

(XLSX)

S2 Table. Matrix of pairwise comparisons made by experts in East African livestock urine exposure during analytic hierarchy process, and calculation of livestock urine exposure weights, 2015.

https://doi.org/10.1371/journal.pntd.0006372.s003

(XLSX)

S3 Table. Matrix of pairwise comparisons made by experts in rodent urine exposure in East Africa during analytic hierarchy process, and calculation of rodent urine exposure weights, 2015.

https://doi.org/10.1371/journal.pntd.0006372.s004

(XLSX)

S4 Table. Matrix of pairwise comparisons made by experts in surface water exposure in East Africa during analytic hierarchy process, and calculation of surface water exposure weights, 2015.

https://doi.org/10.1371/journal.pntd.0006372.s005

(XLSX)

S5 Table. Interaction terms assessed during logistic regression analyses of risk factors for acute leptospirosis and Leptospira seropositivity in northern Tanzania, 2012–14.

https://doi.org/10.1371/journal.pntd.0006372.s006

(DOCX)

S6 Table. Results tables for logistic regression of risk factors for seropositivity to Leptospira serogroup Icterohaemorrhagiae among patients with febrile illness in northern Tanzania.

https://doi.org/10.1371/journal.pntd.0006372.s007

(DOCX)

Acknowledgments

The authors would like to thank those involved in recruitment, laboratory work, data management and study administration, including: Godfrey S. Mushi, Flora W. Mboya, Lilian E. Ngowi, Winfrida H. Shirima, Michael E. Butoyi, Anna H. Mwalla, Miriam L. Barabara, Ephrasia Mariki., Tumsifu G. Tarimo, Yusuf S. Msuya, Leila J. Sawe, Aaron E. Tesha, Luig J. Mbuya, Edward M. Singo, Isaac A. Afwamba, Thomas M. Walongo, Remigi P. Swai, Augustine M. Musyoka, Philoteus A. Sakasaka, O. Michael Omondi, Enoch J. Kessy, Alphonse S. Mushi, Robert S. Chuwa, Cynthia A. Asiyo, Frank M. Kimaro, and Francis P. Karia. In addition we would like to thank the study participants as well as the clinical staff and administration at Kilimanjaro Christian Medical Centre and Mawenzi Regional Referral Hospital for their support during this study and the climate scientists at the Tanzanian Production Company for provision of climate data.

Disclaimer The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. Use of trade names and commercial sources is for identification only and does not imply endorsement by the US Department of Health and Human Services or the Centers for Disease Control and Prevention.

References

  1. 1. Allan KJ, Biggs HM, Halliday JEB, Kazwala RR, Maro VP, Cleaveland S, et al. Epidemiology of leptospirosis in Africa: a systematic review of a neglected zoonosis and a paradigm for 'One Health' in Africa. PLoS Negl Trop Dis. 2015;9(9).
  2. 2. Crump JA, Morrissey AB, Nicholson WL, Massung RF, Stoddard RA, Galloway RL, et al. Etiology of severe non-malaria febrile illness in Northern Tanzania: a prospective cohort study. PLoS Negl Trop Dis. 2013;7(7):e2324. pmid:23875053; PubMed Central PMCID: PMCPMC3715424.
  3. 3. Maze MJ, Biggs HM, Rubach MP, Galloway RL, Cash-Goldwasser S, Allan KJ, et al. Comparison of the estimated incidence of acute leptospirosis in the Kilimanjaro Region of Tanzania between 2007–08 and 2012–14. PLoS Negl Trop Dis. 2016;10(12):e0005165. pmid:27911902; PubMed Central PMCID: PMCPMC5135036.
  4. 4. Mwachui MA, Crump L, Hartskeerl R, Zinsstag J, Hattendorf J. Environmental and behavioural determinants of leptospirosis transmission: a systematic review. PLoS Negl Trop Dis. 2015;9(9):e0003843. pmid:26379035; PubMed Central PMCID: PMCPMC4574979.
  5. 5. Levett PN. Leptospirosis. Clin Microbiol Rev. 2001;14(2):296–326. pmid:11292640
  6. 6. Bovet P, Yersin C, Merien F, Davis CE, Perolat P. Factors associated with clinical leptospirosis: a population-based case-control study in the Seychelles (Indian Ocean). Int J Epidemiol. 1999;28(3):583–90. pmid:10405868.
  7. 7. Sarkar U, Nascimento SF, Barbosa R, Martins R, Nuevo H, Kalofonos I, et al. Population-based case-control investigation of risk factors for leptospirosis during an urban epidemic. Am J Trop Med Hyg. 2002;66(5):605–10. pmid:12201599.
  8. 8. Biggs HM, Bui DM, Galloway RL, Stoddard RA, Shadomy SV, Morrissey AB, et al. Leptospirosis among hospitalized febrile patients in northern Tanzania. Am J Trop Med Hyg. 2011;85(2):275–81. pmid:21813847; PubMed Central PMCID: PMCPMC3144825.
  9. 9. Schoonman L, Swai ES. Risk factors associated with the seroprevalence of leptospirosis, amongst at-risk groups in and around Tanga city, Tanzania. Ann Trop Med Parasitol. 2009;103(8):711–8. pmid:20030995.
  10. 10. Schoonman L, Swai ES. Herd and animal-level risk factors for bovine leptospirosis in Tanga region of Tanzania. Trop Anim Health Prod. 2010;42(7):1565–72. pmid:20517645.
  11. 11. Swai ES, Schoonman L. A survey of zoonotic diseases in trade cattle slaughtered at Tanga city abattoir: a cause of public health concern. Asian Pac J Trop Biomed. 2012;2(1):55–60. pmid:23569835; PubMed Central PMCID: PMCPMC3609208.
  12. 12. Allan KJ. Leptospirosis in northern Tanzania: investigating the role of rodents and ruminant livestock in a neglected public health problem [Doctoral Thesis]. Glasgow: University of Glasgow; 2016.
  13. 13. Ashford DA, Kaiser RM, Spiegel RA, Perkins BA, Weyant RS, Bragg SL, et al. Asymptomatic infection and risk factors for leptospirosis in Nicaragua. Am J Trop Med Hyg. 2000;63(5–6):249–54. pmid:11421372.
  14. 14. Leal-Castellanos CB, Garcia-Suarez R, Gonzalez-Figueroa E, Fuentes-Allen JL, Escobedo-de la Penal J. Risk factors and the prevalence of leptospirosis infection in a rural community of Chiapas, Mexico. Epidemiol Infect. 2003;131(3):1149–56. pmid:14959783; PubMed Central PMCID: PMCPMC2870065.
  15. 15. Sugunan AP, Vijayachari P, Sharma S, Roy S, Manickam P, Natarajaseenivasan K, et al. Risk factors associated with leptospirosis during an outbreak in Middle Andaman, India. Indian J Med Res. 2009;130(1):67–73. pmid:19700804.
  16. 16. Goris MG, Hartskeerl RA. Leptospirosis serodiagnosis by the microscopic agglutination test. Curr Protoc Microbiol. 2014;32:Unit 12E 5. pmid:24510846.
  17. 17. Centers for Disease Control and Prevention. Leptospirosis (Leptospira interrogans) 2013 case definition 2013 [21 October 2015]. Available from: https://wwwn.cdc.gov/nndss/conditions/leptospirosis/case-definition/2013/.
  18. 18. 2012 Population and Housing Census. Dar es Salaam: National Bureau of Statistics (Tanzania), 2013.
  19. 19. Moraga P, Cano J, Baggaley RF, Gyapong JO, Njenga SM, Nikolay B, et al. Modelling the distribution and transmission intensity of lymphatic filariasis in sub-Saharan Africa prior to scaling up interventions: integrated use of geostatistical and mathematical modelling. Parasit Vectors. 2015;8:560. pmid:26496983
  20. 20. International Livestock Research Institute. International Livestock Research Institute annual report 2002: International Livestock Research Institute; 2003.
  21. 21. Land use and vegetation data: National Geomatic Center of China report. Beijing, China: 2010.
  22. 22. Kulldorff M, Nagarwalla N. Spatial disease cluster: Detection and Inference. Statistics in Medicine. 1995;14:799–810. pmid:7644860
  23. 23. Saaty TL. Decision making with the analytic hierarchy process. Int J Services Sciences. 2008;1(1):83–98.
  24. 24. Pecchia L, Bath PA, Pendleton N, Bracale M. Analytic Hierarchy Process (AHP) for examining healthcare professionals' assessments of risk factors. The relative importance of risk factors for falls in community-dwelling older people. Methods Inf Med. 2011;50(5):435–44. pmid:21132218.
  25. 25. Royston P, Ambler G, Sauerbrei W. The use of fractional polynomials to model continuous risk variables in epidemiology. Int J Epidemiol. 1999;28(5):964–74. pmid:10597998.
  26. 26. Akaike H. A new look at the statistical model identification. IEEE Transaction on Automatic Control. 1974;19:716–23.
  27. 27. Cook EA, de Glanville WA, Thomas LF, Kariuki S, Bronsvoort BM, Fevre EM. Risk factors for leptospirosis seropositivity in slaughterhouse workers in western Kenya. Occup Environ Med. 2017;74(5):357–65. pmid:27913579; PubMed Central PMCID: PMCPMC5520261.
  28. 28. Dreyfus A, Dyal JW, Pearson R, Kankya C, Kajura C, Alinaitwe L, et al. Leptospira seroprevalence and risk factors in health centre patients in Hoima District, Western Uganda. PLoS Negl Trop Dis. 2016;10(8):e0004858. pmid:27487398.
  29. 29. Victoriano AF, Smythe LD, Gloriani-Barzaga N, Cavinta LL, Kasai T, Limpakarnjanarat K, et al. Leptospirosis in the Asia Pacific region. BMC Infect Dis. 2009;9:147. pmid:19732423; PubMed Central PMCID: PMCPMC2749047.
  30. 30. Sakundarno M, Bertolatti D, Maycock B, Spickett J, Dhaliwal S. Risk factors for leptospirosis infection in humans and implications for public health intervention in Indonesia and the Asia-Pacific region. Asia Pac J Public Health. 2014;26(1):15–32. pmid:24097928.
  31. 31. Allan KJ, Halliday JEB, Moseley M, Carter RW, Ahmed A, Goris MGA et al. Assessment of animal hosts of pathogenic Leptospira in northern Tanzania. PLoS Negl Trop Dis. in press
  32. 32. Rice production in sub-Sahara Africa [8 August 2016]. Available from: http://riceforafrica.net/index.php/production.
  33. 33. Kawaguchi L, Sengkeopraseuth B, Tsuyuoka R, Koizumi N, Akashi H, Vongphrachanh P, et al. Seroprevalence of leptospirosis and risk factor analysis in flood-prone rural areas in Lao PDR. Am J Trop Med Hyg. 2008;78(6):957–61. pmid:18541776.
  34. 34. Lau CL, Watson CH, Lowry JH, David MC, Craig SB, Wynwood SJ, et al. Human leptospirosis infection in Fiji: an eco-epidemiological approach to identifying risk factors and environmental drivers for transmission. PLoS Negl Trop Dis. 2016;10(1):e0004405. pmid:26820752; PubMed Central PMCID: PMCPMC4731082.
  35. 35. Zhang HL, Mnzava KW, Mitchell ST, Melubo ML, Kibona TJ, Cleaveland S, et al. Mixed methods survey of zoonotic disease awareness and practice among animal and human healthcare providers in Moshi, Tanzania. PLoS Negl Trop Dis. 2016;10(3):e0004476. pmid:26943334.
  36. 36. Mgode GF, Machang'u RS, Goris MG, Engelbert M, Sondij S, Hartskeerl RA. New Leptospira serovar Sokoine of serogroup Icterohaemorrhagiae from cattle in Tanzania. Int J Syst Evol Microbiol. 2006;56(Pt 3):593–7. pmid:16514033.
  37. 37. Ganoza CA, Matthias MA, Collins-Richards D, Brouwer KC, Cunningham CB, Segura ER, et al. Determining risk for severe leptospirosis by molecular analysis of environmental surface waters for pathogenic Leptospira. PLoS Med. 2006;3(8):e308. pmid:16933963; PubMed Central PMCID: PMCPMC1551915.
  38. 38. Levett PN. Usefulness of serologic analysis as a predictor of the infecting serovar in patients with severe leptospirosis. Clin Infect Dis. 2003;36(4):447–52. pmid:12567302.
  39. 39. Limmathurotsakul D, Turner EL, Wuthiekanun V, Thaipadungpanit J, Suputtamongkol Y, Chierakul W, et al. Fool's gold: Why imperfect reference tests are undermining the evaluation of novel diagnostics: a reevaluation of 5 diagnostic tests for leptospirosis. Clin Infect Dis. 2012;55(3):322–31. pmid:22523263; PubMed Central PMCID: PMCPMC3393707.
  40. 40. Goris MG, Leeflang MM, Boer KR, Goeijenbier M, van Gorp EC, Wagenaar JF, et al. Establishment of valid laboratory case definition for human leptospirosis. Journal of Bacteriology & Parasitology. 2012;2012.
  41. 41. Cumberland P, Everard CO, Wheeler JG, Levett PN. Persistence of anti-leptospiral IgM, IgG and agglutinating antibodies in patients presenting with acute febrile illness in Barbados 1979–1989. Eur J Epidemiol. 2001;17(7):601–8. pmid:12086073.
  42. 42. Maze MJ, Galloway RL, Rubach MP, Allan KJ, Biggs HM, Halliday JE, et al., editors. Three Species Of Leptospira Detected In Blood Of Patients With Fever In Northern Tanzania. 10th International Leptospirosis Society Meeting; 2017 27 Nov-1 Dec 2017; Palmerston North, New Zealand.
  43. 43. Maze MJ, Rubach MP, Sakasaka P, Swai R, Afwamba I, Galloway RL, et al. Evaluation of performance of three commercially available leptospirosis IgM point-of-care tests to diagnose acute leptospirosis among patients with fever in northern Tanzania. 10th International Leptospirosis Society Meeting; Palmerston North, New Zealand. 2017.
  44. 44. Schreier S, Doungchawee G, Chadsuthi S, Triampo D, Triampo W. Leptospirosis: current situation and trends of specific laboratory tests. Expert Rev Clin Immunol. 2013;9(3):263–80. pmid:23445200.
  45. 45. Picardeau M, Bertherat E, Jancloes M, Skouloudis AN, Durski K, Hartskeerl RA. Rapid tests for diagnosis of leptospirosis: current tools and emerging technologies. Diagn Microbiol Infect Dis. 2014;78(1):1–8. pmid:24207075.
  46. 46. Gonzalez S, Geymonat JP, Hernandez E, Marques JM, Schelotto F, Varela G. Usefulness of real-time PCR assay targeting LipL32 gene for diagnosis of human leptospirosis in Uruguay. J. 2013;7(12):941–5. https://dx.doi.org/10.3855/jidc.4110. pmid:24334940.
  47. 47. Murray CK, Gray MR, Mende K, Parker TM, Samir A, Rahman BA, et al. Use of patient-specific Leptospira isolates in the diagnosis of leptospirosis employing microscopic agglutination testing (MAT). Trans R Soc Trop Med Hyg. 2011;105(4):209–13. pmid:21334705.
  48. 48. Smythe LD, Wuthiekanun V, Chierakul W, Suputtamongkol Y, Tiengrim S, Dohnt MF, et al. The microscopic agglutination test (MAT) is an unreliable predictor of infecting Leptospira serovar in Thailand. Am J Trop Med Hyg. 2009;81(4):695–7. pmid:19815889.
  49. 49. Hartskeerl RA, Collares-Pereira M, Ellis WA. Emergence, control and re-emerging leptospirosis: dynamics of infection in the changing world. Clin Microbiol Infect. 2011;17(4):494–501. pmid:21414083.