Skip to main content
NCBI home page
Transactions of the Royal Society of Tropical Medicine and Hygiene logoLink to Transactions of the Royal Society of Tropical Medicine and Hygiene
. 2024 Apr 1;118(9):605–615. doi: 10.1093/trstmh/trae013

Risk of human leptospirosis in Colombia: spatiotemporal analysis and related hydroclimatic factors

Juan José Montenegro-Idrogo 1,✉, D Katterine Bonilla-Aldana 2, Alfonso J Rodríguez-Morales 3,4,5
PMCID: PMC13397013  PMID: 38557833

Abstract

Background

Leptospirosis is an endemic zoonosis in tropical areas that is mainly related to rural activities; nevertheless, human leptospirosis (HL) outbreaks differ among regions. In Colombia, HL notifications are mandatory. Our objective was to determine the spatiotemporal distribution of HL in Colombia during 2007–2018 and its relationship with the main hydroclimatic variables.

Methods

We determined the estimated incidence and lethality of HL according to department and year. The Bayesian spatiotemporal analysis of an autoregressive model (STAR) model included HL cases and hydroclimatic factors (average temperature, rainfall and relative humidity) for quarterly periods.

Results

During the study period, 10 586 HL cases were registered (estimated incidence: 1.75 cases x 105) and 243 deaths by HL (lethality 2.3%). The STAR model found association of HL risk with temperature (RR:6.80; 95% CI 3.57 to 12.48) and space. Quindío and three other Amazonian departments (Guainía, Guaviare and Putumayo) had a positive relationship with a significant number of HL cases, adjusted for quarterly precipitation and humidity.

Conclusion

Spatial analysis showed a high risk of HL in departments of the western Andean Colombian regions. By contrast, in the spatiotemporal model, a higher HL risk was associated with temperature and departments of the North Colombian Amazon regions and Quindío in the Colombian Andean region.

Keywords: disease clusters, epidemiology, leptospirosis, spatiotemporal analysis

Introduction

Human leptospirosis (HL) is caused by bacteria of the genus Leptospira spp., mainly Leptospira interrogans and its different variants.1–3 In Colombia, the most frequently described serovars were Australis (22.4%), followed by Ballum (10.4%) and Autumnalis (9.0%). Australis was more frequently reported in severe HL cases and deaths due to HL. However, there was no statistical difference in comparison with other serovars for these severity variables during 2015–2020.2 The highest morbidity and mortality estimates attributed to HL are described in tropical regions such as Oceania, Southeast Asia and the Caribbean (presenting morbidity of 150, 55 and 51 cases per 105 inhabitants and mortality of 9.6, 3 and 2.9 deaths per 105 inhabitants, respectively), followed by the Eastern sub-Saharan Africa and Andean regions in Latin America (morbidity of 26 and 22 cases per 105 inhabitants; and mortality of 1.9 and 0.9 deaths per 105 inhabitants, respectively).3 In the latter region, Colombia is considered one of the highest endemic areas of HL and cases with acute febrile disease are subject to notification and epidemiologic surveillance.1,4,5 Thus, in endemic areas, the high number of severe cases and mortality by HL are likely related to the larger number of cases.3,6

Spatiotemporal studies on HL performed from 1998 to 2017 were mainly carried out in Brazil, the USA, Canada, Samoa and Thailand,7–11 with models being developed during this period to identify patterns related to the occurrence of HL outbreaks.7,9 Some studies have suggested an association between HL and flooding and rural farming or non-farming activities.12 By contrast, others have suggested that changes in HL outbreaks were related to alterations in temperature, relative humidity, rainfall or solar radiation.13 In Latin America, these studies were carried out in Brazil10,14–17 and described seasonal trends related to some sociodemographic determinants. In Argentina, the Oceanic Niño Index was postulated to predict outbreaks in a district analysis of its northeastern regions.14

In Colombia, some departments, such as Antioquia and Córdoba, reported the highest frequency of HL,4,18,19 with an increase in HL cases of 25% in 2010–2011 during the La Niña phenomenon, and a reduction of 17% in HL cases during the El Niño phenomenon at the country level. However, in endemic areas, an increasing trend has been reported according to department.20 This finding suggests likely seasonal behavior.5,21 Additionally, an increase in HL cases was reported in municipalities, with greater areas dedicated to agricultural activities in the ‘Coffee triangle’.22 Nonetheless, the association between outbreaks of HL cases and seasonality or sociodemographic determinants remains unclear.5,18,19 This is probably due to the heterogeneity of the hydroclimatic behavior in different Colombian regions, hindering the recognition of behaviors or trends that allow determining how these factors interact to predict the complex behavior in an HL outbreak in each Colombian area.

Climatic changes in recent decades could affect the behavior of some climatic variables (temperature, rainfall and humidity, etc.) that may experience significant changes in different time periods, which could reinforce other determinants of disease in each geographical area.1 This translates into the increase/emergence of endemic infectious diseases such as leptospirosis (and others such as dengue, malaria and diarrhea outbreaks), which may also be affected, depending on the region. A spatiotemporal analysis allows the identification of the interaction between areas and specific periods adjusted for seasonal variables to identify regions with probable disease risk,7 as opposed to conventional analyses such as prevalence/incidence studies.

Our study aimed to determine the spatiotemporal distribution of HL cases in Colombia over a period of 11 y and describe any association with hydroclimatic factors.

Materials and Methods

Design

This was an ecological study and spatiotemporal analysis based on all notified cases of HL in Colombian departments and municipalities during 2007–2018 distributed into quarterly periods.

Study location and population

Colombia is a Latin American country located in northern South America (latitude 4.5709º N, longitude 74.2973º W). The country is divided into six geographical regions: the Caribbean, Pacific and Insular regions (with sea territory); and the Andean region, Orinoquía and Amazon regions, with the latter two having the highest tropical and humid climates. At the political level, Colombia is divided into 1123 municipalities with 32 departments. The estimated population in 2021 was 51 049 498 habitants with a density of 42.3 inhabitants/sq. km. The highest density is located in the western regions (Figure 1).

Figure 1.

Figure 1.

Open in a new tab

Colombia map distributed by departments and regions.

Source of information and procedures

The unit of analysis was the estimated yearly and quarterly distribution of each study variable at the departmental level. Data were obtained from the following databases:

  1. SIVIGILA (https://portalsivigila.ins.gov.co/), the epidemiological surveillance system in Colombia, provided information about notified HL cases and the estimated annual population. With this information we calculated the estimated incidence (number of notified HL cases/estimated population x 105 inhab), lethality (number of deaths with HL/total notified HL cases) and hospitalization for HL (number of HL hospitalized/total notified HL cases).

  2. The Prediction of Worldwide Energy Resources project (POWER) database (https://power.larc.nasa.gov/data-access-viewer/). This is a georeferenced NASA (National Aeronautics and Space Administration) information system that allows for the recovery of hydroclimatic variables (average temperature, relative humidity and average rainfall) in quarterly periods for each geographical area at the world level.

Case definition of HL

A confirmed laboratory case was defined according to the Colombian Surveillance of Leptospirosis protocol,23 in which titers above the cut-off of the microscopic agglutination test indicate recent infection in suspected cases in which PCR or positive cultures are not available.

All the databases are open access and the data were collected in *csv format and included in an Excel file after checking for inconsistencies three times. The final file, together with a file in *shp format that includes the political distribution of Colombia up to the district level, was included for analysis in the statistical package R (2009–2024 Posit Software, PBC). The final data were expressed in quarterly and yearly periods for each analysis in each department and district.

Analysis with sociodemographic variables

The main characteristics of HL were described as well as the reported cases and the estimated incidence, number of cases, lethality and hospitalization for HL.

Spatiotemporal analysis

The spatiotemporal analysis was performed on a sequential process in three phases: (i) temporal component assessment (observed cases, expected cases and standardized case ratio, quarterly in Colombia) and the average hydroclimatic values (average temperature, relative humidity, average rainfalls); (ii) spatial component assessment according to total cases observed and standardized case ratio of HL during the study period (2007–2018) by the Colombian department; and (iii) spatiotemporal analysis of observed HL cases, including the two previous components adjusted to previous hydroclimatic variables.

In the temporal component analysis, HL cases were assessed as observed and by the standardized case ratio (observed/expected). The assumption of stationarity was made using the Dickey–Fuller test, and the random walk (RW) assumption was assessed using the Ljung–Box test. The spatial component was described by a quarter during the study period for HL cases as observed and the standardized case ratio in exploratory graphic maps. The effect was considered statistically significant according to the Moran local index to identify potential clusters and stratification in hot (red) and cold (blue) zones, which indicated high and low values, respectively.

Spatiotemporal analysis was performed using a structured additive regression (spatiotemporal analysis autoregressive [STAR]) model, a Bayesian approximation of generalized regression models to determine the presence of HL clusters, which considered the spatial effect (departmental level) and temporal effect (quarterly period) and was also adjusted for hydroclimatic variables (average temperature, relative humidity and average rainfall). This model considered that HL case numbers complied with the RW assumption and intrinsic autoregressive conditional assumption.

This methodology24,25 is based on the analysis of HL cases in a time ‘t’ and in a place ‘i‘ (Oi,t) that complies with a Poisson-gamma distribution of the average number of observed cases in a time ‘t’ and in a place ‘i’ (μ I,t), where μ i,t = Ei, t · θi,t, ‘Ei,t’ is considered to be expected cases for the time ‘t’ and the place ‘i‘, while ‘θi, t’ is the relative risk in the time ‘t’ and the place ‘i’.

The disease occurrence effect observed in a specific place ‘i’ in a time ‘t’ was assessed with the algorithm: log(θi, t) = α + βxi + vi + wt; where

θi,t: relative risk (RR) in the time ‘t’ and space ‘i’.

xi: covariate of interest (average temperature, relative humidity, average rainfall).

vi: spatial effect in the place ‘i’.

wt: temporal effect in the time ‘t’.

We calculated the RR or rate ratio/mean (coefficient of the model) with 95% CIs.

Ethical aspects

This study was approved by the Universidad Científica del Sur IRB (Register Nº 318–2021-POS15).

Results

Incidence of leptospirosis

During the study period, the estimated incidence of HL was 1.75 cases per 105 inhabitants. The global hospitalization rate due to HL was 58.81%, and the global lethality was 2.3%. Notified HL cases were mainly males (65.5%), from rural setting/villages (18.1%) and of an economically active age (20–59 y; 70.3%). The annual distribution of these variables was similar, with the highest number of HL cases reported from 2010 to 2012. The years with the highest percentage of hospitalization and mortality due to HL were 2017 and 2014, respectively (Table 1).

Table 1.

Yearly relative distribution characteristics of HL cases in Colombia (2007–2018)

Characteristic 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Total
Male sex (%) 49.44 58.33 57.73 63.47 66.96 68.47 68.84 70.64 64.35 68.52 72.77 76.62 65.52
Rural origin (%) 13.90 9.20 13.76 10.78 12.23 15.16 12.53 13.56 15.45 15.84 14.70 14.97 13.39
Rural or village origin (%) 16.57 12.88 18.65 14.78 18.34 21.50 18.91 17.25 19.48 21.52 19.07 18.03 18.07
Age
 <20 y 2.95 7.01 7.74 5.19 6.23 9.88 9.82 8.87 13.98 16.31 15.14 18.49 9.70
 20–59 y 71.36 69.82 69.98 71.20 72.17 72.40 70.79 72.29 68.41 66.15 66.97 70.07 70.32
 ≥60 y 25.70 23.17 22.29 23.61 21.60 17.72 19.39 18.84 17.61 17.54 17.89 11.44 19.98
Hospitalized cases (%) 34.41 56.64 48.32 72.08 4.02 68.33 75.29 66.14 74.94 75.78 78.46 76.36 58.81
Lethality (% deaths) 3.51 2.10 1.02 2.78 2.57 2.30 2.78 2.84 1.69 2.24 2.04 1.02 2.30
Total cases 712 761 981 1225 1243 1042 862 951 770 669 687 588 10 586
Incidence (105 inhab) 1.62 1.71 2.18 2.69 2.70 2.24 1.83 2.00 1.60 1.37 1.39 1.18 1.75
Open in a new tab

The departments with the highest estimated incidence of HL were Guaviare (37.58 cases/105 inhabitants [inhab]), Risaralda (6.2 cases/105 inhab), Valle del Cauca (4.19 cases/105 inhab) and Atlántico (3.91 cases/105 inhab). Departments with the greatest number of deaths due to HL were Bolívar (57), Atlántico (36), Valle del Cauca (33) and Magdalena (21), with Bolívar presenting the highest lethality (8.2%). The municipalities with the highest number of reported HL cases (>500 cumulative cases) were Cali (Valle del Cauca), Barranquilla (Atlántico), Apartadó and Medellin (Antioquia), San José del Guaviare (Guaviare) and Cartagena (Bolívar).

Temporal analysis

A peak in the number of HL cases was identified around the last quarter of 2011, with a sustained decline in cases and the standardized ratio thereafter to values <1 in the last quarter (Supplemental Information S1). With regard to hydroclimatic variables, there was an increase in the average values at the country level, with a peak decrease in the first quarter of 2010 and 2016, associated with a reduction in the average rainfall values and an increase in temperature during the same time periods (S2). The seasonality assumption was not met in the initial assessment (Dickey–Fuller test p=0.636); the series was transformed by means of the first difference to comply with this assumption (Dickey–Fuller test p<0.001) (Figure 2). Additionally, we assessed the number of autoregressive terms and the number of moving averages required to achieve validity (white noise or RW), and the outcomes obtained by the autocorrelation function and partial autocorrelation function were validated (S3). A non-significant Ljung–Box test (p=0.776) determined the RW.

Figure 2.

Figure 2.

Open in a new tab

Time-series quarterly distribution of reported cases of human leptospirosis in Colombia, 2007–2018, transformed with first difference.

Spatial analysis

We described the cumulative departmental distribution of reported cases, expected cases and standardized case ratio; we worked with the latter to reduce the bias in the spatial analysis. The cumulative analysis showed an increased frequency of cases in the middle-western region of Colombia, mainly around the coffee-growing and Caribbean regions (Figure 3). At the departmental level, the Moran local index only identified one hotspot (a cluster with high values) for reported cases in Risaralda (p<0.01) (Figure 4A); for the standardized case ratio level, the Moran local index identified one cold spot (a cluster with low values) around the middle of the country (Cundinamarca, Boyacá and Casanare), which was of greatest relevance in Casanare (p<0.001) (Figure 4B). In relation to the quarterly distribution of reported cases, there was an increase in cases in the Tolima department (>50 cases per quarter) in 21 of the 48 time periods assessed, including the third quarter of 2009, during which >200 cases were reported (Figure 5). In the analysis of the standardized case ratio by quarters, the Vichada department had a standardized ratio of cases >10 in 26 of the 48 time periods assessed (Figure 6).

Figure 3.

Figure 3.

Open in a new tab

Consolidated cases: (A) observed; (B) expected; and (C) standardized ratio of cases of human leptospirosis in Colombia, 2007–2018.

Figure 4.

Figure 4.

Open in a new tab

Moran local index in Colombian spatial analysis (blue/red) and statistical significance (green) (A) in total observed cases; and (B) according to standardized ratio of cases.

Figure 5.

Figure 5.

Open in a new tab

Quarterly distribution of reported cases of human leptospirosis by department in Colombia from 2007 to 2018.

Figure 6.

Figure 6.

Open in a new tab

Quarterly distribution of standardized ratio of cases of human leptospirosis by department in Colombia during 2007–2018. *Standardized ratio of cases of human leptospirosis (SRHL)=observed HL cases/expected HL cases.

Spatiotemporal analysis

The STAR model identified an association between the number of HL cases and the departments (RR:6.16; 95% CI 3.53 to 9.93). Quindío, Guainía, Guaviare and Putumayo were hotspots (Figure 7A) with a positive relationship with the number of HL cases. In the adjusted model, the number of HL cases showed a decreasing trend. The three hydroclimatic parameters included in the model did not show seasonality (statistically significant Dickey–Fuller test), leading us to suspect that the main component of variability was longitudinal. The graphical analysis for each covariate smoothed at 20 knots, showing that temperature and rainfall had a positive relationship, while only temperature was statistically associated in the adjusted model (RR: 6.80; 95% CI 3.57 to 12.48), which also included humidity (RR: 2.34; 95% CI 0.74 to 6.17) and rainfall (RR: 0.90; 95% CI 0.34 to 2.06) (Figure 7).

Figure 7.

Figure 7.

Open in a new tab

Spatiotemporal analysis, STAR model and graphical smoothed estimates with confidence intervals of covariates included in the model in Colombia, 2007–2018. (A) STAR model showing hotspot (red) and cold spot (blue); smoothed estimates with confidence intervals of covariates included in the STAR model: (B) number of cases; (C) humidity; (D) rainfall; and (E) temperature.

Discussion

In 2014, Colombia was one of the Latin American countries that reported a high incidence of HL, after Brazil, Peru and Ecuador. The HL in Colombia is notifiable and has been circumscribed to a Public Health plan for 10 y (2012–2021).6

Before this period, epidemiological studies described a prevalence of HL of <10% in zones such as Tolima, Atlántico and Boyacá, with the highest prevalence of >30% being reported in Antioquia and Córdoba. In addition, the prevalence of reported animal leptospirosis is greater in these areas (in dogs, rodents and primates).18 During 2010 and 2011, an increase in the number of cases of 25% was reported at the country level during the ‘La Niña’ phenomenon,19,20 while a decrease of 17% was observed during the ‘El Niño’ phenomenon. However, at the department level, there was an increase in notified cases in endemic zones during both the phenomenon periods.20 These variations in specific periods are concordant with the difference between the findings of the initial spatial analysis by department and those found in our spatiotemporal analysis, including the effects of the estimated hydroclimatic variables.

Schneider et al.6 from a regional Latin American data analysis describe to 2014 some characteristics related to HL cases in Colombia, such as the frequency of HL cases in a primarily urban setting (74.6%) with a predominantly male population (70.9%), with the main factor related to HL being contact with animals. These findings are similar to those found during our study period (2007–2018), in which the overall percentage of people with HL of non-rural origin was >80% (including populated centers as a rural area) and the overall percentage of males with HL was 65.52%, with a trend toward an increase of >70% in recent years (Table 1). In other regions of South America, such as Brazil (Santa Catarina 2001–2015),26 the risk of HL cases was described mainly in urban areas and was related to periods coinciding with natural disasters due to heavy rains or floods, during which outbreak peaks were preceded by periods of increased rainfall. Other global Brazilian analyses with annual periodicity27 determined clusters of greater association in the central-western and specific southern parts of Brazil (the latter being described by Teles et al.28), in addition to other clusters in the northern region, with these municipalities having presented at least one flooding event as a common characteristic. It is of note that the phenomena of increased rainfall or flooding is related to increased HL outbreaks in tropical areas of South America. This circumstance has also been described in other regions such as Thailand.29 However, our spatiotemporal model in Colombia did not show this occurrence, but it did determine areas near Brazil in the Amazon region with a high association of HL. Another analysis in the Brazilian state of Sergipe (2008–2017)30 also showed an association of HL with urban areas and also compared variables related to increased social inequities in this state, showing similar outcomes to those found in Rio Grande do Sul, Brazil (2010–2019).31 It should be noted that in this last study the periods were annualized, thereby reducing the probability of evaluating the seasonal effect by periodic hydroclimatic variables.

García-Ramírez et al. evaluated municipalities of the Colombia Coffee Triangle22 to identify occupational determinants related to the estimated HL incidence in 53 municipalities of this region in the 2007–2011 period and repercussion with farming areas. The areas with the highest incidence of HL were Pueblo Rico (Risaralda) followed by Armenia (Quindío). There was a statistically significant association between the number of HL cases in 2008 and the proportion of harvested areas in each municipality. Regarding these findings, it should be noted that Risaralda is one of the departments that showed a statistically significant spatial association (hotspot) (Figure 4A); however, Quindío presented a significant positive effect in the spatiotemporal analysis in the STAR model, unlike in the other regions of the Colombian coffee-growing area. It was of note that this department was shown to be a potential endemic focus of HL in Colombia, even in the analysis adjusted for the estimated hydroclimatic variables (Figures 4A and 7).

In a Colombian temporal series analysis of 2007–2016, Gutierrez32 assessed trends of HL with meteorological factors (rainfall and average temperature) distributed over 0–4-wk periods using a negative binomial regression multilevel method to determine this behavior in 30 municipalities. The study described a positive relationship between rainfall and increased HL in weekly periods, adjusting only for these two meteorological components, and there was no evidence of seasonal cycles with HL outbreaks or adjustment for other variables previously described in this period. Our analysis was performed on a quarterly basis and at the departmental level by means of a Bayesian STAR method, which showed a statistical association with the variation in temperature; thus, an increase of one unit of degree Celsius increased the number of HL by 6.6-fold, adjusted to departments, humidity and precipitation (the latter without statistical association). It is understood that in different ways, this association is related to departments, with some departments having a positive association with HL cases and others having a negative association with the same.

In the same period, Gutiérrez et al.19 conducted an ecological study to analyze the environmental and sociodemographic variables that could influence the presentation of HL cases. They identified four critical HL points, which included 18 municipalities, using a negative binomial regression multiple method. Each deleted variable was revalued in a parsimonious multivariate model, and variables with high correlation were tested by interchanging in that model. Finally, socioeconomic determinants such as education level, poverty, social marginalization related to violence and ecological variables (rainfall, altitude and forest coverage) could have been related to the incidence of HL in this period. It is evident that the number of HL cases reported has some relationship with areas of higher population density, such as areas corresponding to the Colombian coffee-growing triangle (including Quindio, Caldas and Risaralda departments). However, spatiotemporal analysis did not necessarily show an association with any of these regions, with only some departments being considered as potential foci of HL outbreaks, such as Quindío. The remaining departments may have a greater number of cases due to an increase in population density in these areas, while Nariño, Caldas, Magdalena and Bolívar could be considered departments with a low risk of outbreaks. In our quarterly distribution series, we found no evidence of a statistical association between rainfall or humidity, at least in the spatiotemporal model of Colombia during the study period.

Another analysis by Gutiérrez and Martínez-Vega during the same period (2007–2016)33 identified six clusters (five in the Andean region and one in the Caribbean region). El Niño Southern Oscillation (ENSO) Cycle and rainfall anomalies were assessed, as well as the relevance of the La Niña phenomenon, with excess rainfall periods and the occurrence of HL cases. The clusters were identified using a standardized Monte Carlo method considering p<0.01. A permutation spatiotemporal model is used in this study. The cases observed for each group were compared with the expected cases for each municipality (proportional to its population size) if the spatial and temporal locations in all cases were independent of each other, so that there was no space–time interaction (in municipalities with the highest proportion of HL cases). Additionally, these authors suggested the relevant role of flood-prone areas and valleys as potential zones with a major HL risk.

With climate change, behavioral variations occur at the ecological and population levels that can be maintained over time, or in some areas predisposed to these changes, where new patterns of disease behaviors that cannot be identified by descriptive statistics of outbreaks could emerge. At the public health level, the identification of these patterns would allow early intervention in high-risk scenarios when predisposing conditions may appear. In the case of leptospirosis, climatic changes and variations in the patterns of hydroclimatic factors may predispose to changes in some determinants of the disease in endemic areas with a high risk of outbreaks.34,35

Conclusion

We can conclude that, from 2007 to 2018, the cumulative incidence of HL was higher in the departments of Guaviare, Risaralda, Valle del Cauca and Atlántico; however, the HL risk pattern with spatiotemporal analysis adjusted for temperature, rainfall and humidity identified those departments with the highest HL risk (i.e. hotspots): the northern Colombian Amazonia region (Guainía, Guaviare and Putumayo) and Quindío separately, in contrast to departments with the lowest HL risk (i.e. cold spots): Nariño, Magdalena and Bolívar (in western Colombia) and Caldas separately. An increase in HL cases was associated with an increase in temperature in the adjusted model.

Supplementary Material

trae013_Supplemental_File
trae013_supplemental_file.docx (1.7MB, docx)

Contributor Information

Juan José Montenegro-Idrogo, Facultad de Ciencias de la Salud, Universidad Científica del Sur, Lima 150142, Perú.

D Katterine Bonilla-Aldana, Research Unit, Universidad Continental, Huancayo 12001, Perú.

Alfonso J Rodríguez-Morales, Facultad de Ciencias de la Salud, Universidad Científica del Sur, Lima 150142, Perú; Grupo de Investigación Biomedicina, Facultad de Medicina, Fundación Universitaria Autónoma de las Américas, Pereira 660005, Colombia; Gilbert and Rose-Marie Chagoury School of Medicine, Lebanese American University, P.O. Box 13-5053, Beirut, Lebanon.

Authors’ contributions

All the authors conceived the study; and designed the study protocol and carried out the data analysis and interpretation; JMI drafted the manuscript; all the authors critically revised the manuscript for intellectual content and read and approved the final manuscript.

Funding

The Universidad Científica del Sur supported this work: ‘Beca Cabieses’ for postgraduate thesis 2022-1 [RD 019-DGIDI-CIENTIFICA-2021] to JMI.

Competing interests

The authors declare no competing interests relevant to this article's content.

Ethical approval

This study was approved by The Institutional Ethics and Research Committee of the Universidad Científica del Sur (CIEI- CIENTIFICA) (318-2021-POS15).

Data availability

None.

References

  • 1. Lau  CL, Smythe  LD, Craig  SB  et al.  Climate change, flooding, urbanisation and leptospirosis: fuelling the fire?  Trans R Soc Trop Med Hyg. 2010;104(4):631–8. [DOI] [PubMed] [Google Scholar]
  • 2. Parra Barrera  EL, Bello Piruccini  S, Rodríguez  K  et al.  Demographic and clinical risk factors associated with severity of lab-confirmed human leptospirosis in Colombia, 2015-2020. PLoS Negl Trop Dis. 2023;17(7):e0011454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Costa  F, Hagan  JE, Calcagno  J  et al.  Global morbidity and mortality of leptospirosis: a systematic review. PLoS Negl Trop Dis. 2015;9(9):e0003898. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Carreño  LA, Salas  D, Beltrán  KB.  Prevalencia de Leptospirosis en Colombia: revisión sistemática de literatura [Prevalence of leptospirosis in Colombia: systematic literature review]. Rev Salud Publica. 2017;19(2):204–9. [DOI] [PubMed] [Google Scholar]
  • 5. Ramírez-Ramírez  MM, León-Castañeda  OM, Rodriguez-Morales  AJ.  Leptospirosis in an urban setting: cases diagnosed at a private medical center of Western Colombia, 2008-2012. Recent Pat Antiinfect Drug Discov. 2015;10(1):59–63. [DOI] [PubMed] [Google Scholar]
  • 6. Schneider  MC, Leonel  DG, Hamrick  PN  et al.  Leptospirosis in Latin America: exploring the first set of regional data. Rev Panam Salud Publica. 2017;41:e81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Souza  IPO, Uberti  MS, Tassinari  WS.  Geoprocessing and spatial analysis for identifying leptospirosis risk areas: a systematic review. Rev Inst Med Trop Sao Paulo. 2020;62:e35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Narkkul  U, Thaipadungpanit  J, Srisawat  N  et al.  Human, animal, water source interactions and leptospirosis in Thailand. Sci Rep. 2021;11(1):3215. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Suwanpakdee  S, Kaewkungwal  J, White  LJ  et al.  Spatio-temporal patterns of leptospirosis in Thailand: is flooding a risk factor?  Epidemiol Infect. 2015;143(10):2106–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Guimarães  RM, Cruz  OG, Parreira  VG  et al.  Análise temporal da relação entre leptospirose e ocorrência de inundações por chuvas no município do Rio de Janeiro, Brasil, 2007-2012 [Temporal analysis of the relationship between leptospirosis and the occurrence of flooding due to rainfall in the city of Rio de Janeiro, Brazil, 2007-2012]. Cien Saude Colet. 2014;19(9):3683–92. [DOI] [PubMed] [Google Scholar]
  • 11. Joshi  YP, Kim  EH, Cheong  HK.  The influence of climatic factors on the development of hemorrhagic fever with renal syndrome and leptospirosis during the peak season in Korea: an ecologic study. BMC Infect Dis. 2017;17(1):406. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Levy  K, Smith  SM, Carlton  EJ.  Climate change impacts on waterborne diseases: moving toward designing interventions. Curr Environ Health Rep. 2018;5(2):272–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Miller  E, Barragan  V, Chiriboga  J  et al.  Leptospira in river and soil in a highly endemic area of Ecuador. BMC Microbiol. 2021;21(1):17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. López  MS, Müller  GV, Lovino  MA  et al.  Aragonés Pomares L. Spatio-temporal analysis of leptospirosis incidence and its relationship with hydroclimatic indicators in northeastern Argentina. Sci Total Environ. 2019;694:133651. [DOI] [PubMed] [Google Scholar]
  • 15. Gonçalves  NV, Araujo  EN, Sousa  AD  Júnior  et al.  Distribuição espaço-temporal da leptospirose e fatores de risco em Belém, Pará, Brasil [Leptospirosis space-time distribution and risk factors in Belém, Pará, Brazil]. Cien Saude Colet. 2016;21(12):3947–55. [DOI] [PubMed] [Google Scholar]
  • 16. Cerveira  RA, Ferreira  LO, Oliveira  EF  et al.  Spatio-temporal analysis of leptospirosis in Eastern Amazon, State of Pará, Brazil. Rev Bras Epidemiol. 2020;23:e200041. [DOI] [PubMed] [Google Scholar]
  • 17. Baquero  OS, Machado  G.  Spatiotemporal dynamics and risk factors for human leptospirosis in Brazil [published correction appears in Sci Rep. 2019 May 1;9(1):7000]. Sci Rep. 2018;8(1):15170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Dechner  A.  A retrospective analysis of the leptospirosis research in Colombia. J Infect Dev Ctries. 2014;8(3):258–64. [DOI] [PubMed] [Google Scholar]
  • 19. Gutiérrez  JD, Martínez-Vega  RA, Botello  H  et al.  Environmental and socioeconomic determinants of leptospirosis incidence in Colombia [published correction appears in Cad Saude Publica. 2019 May 20;35(5): eER118417]. Cad Saude Publica. 2019;35(3):e00118417. [DOI] [PubMed] [Google Scholar]
  • 20. Arias-Monsalve  C, Builes-Jaramillo  A.  Impact of El Niño-southern oscillation on human leptospirosis in Colombia at different spatial scales. J Infect Dev Ctries. 2019;13(12):1108–16. [DOI] [PubMed] [Google Scholar]
  • 21. Cardona-Ospina  JA, Jiménez-Canizales  CE, Vásquez-Serna  H  et al.  Fatal dengue, chikungunya and leptospirosis: the importance of assessing co-infections in febrile patients in tropical areas. Trop Med Infect Dis. 2018;3(4):123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. García-Ramírez  LM, Giraldo-Pulgarin  JY, Agudelo-Marín  N  et al.  Geographical and occupational aspects of leptospirosis in the coffee-triangle region of Colombia, 2007-2011. Recent Pat Antiinfect Drug Discov. 2015;10(1):42–50. [DOI] [PubMed] [Google Scholar]
  • 23. Protocolo de Vigilancia en Salud Pública de Leptospirosis . versión 1. [Internet]. Instituto Nacional de Salud;2022 ene [cited September 23, 2023]. Available at: https://www.ins.gov.co/buscador-eventos/SitePages/Evento.aspx?Event=48 [accessed 15 March 2024].
  • 24. Bivand  RS, Pebesma  E, Gómez-Rubio  V.  Applied Spatial Data Analysis with R. New York, NY: Springer; 2013. [Google Scholar]
  • 25. Bivand  RS, Pebesma  E, Gómez-Rubio  V.  Spatio-Temporal Data. In:  Bivand  RS, Pebesma  E, Gómez-Rubio  V, editors. Applied Spatial Data Analysis with R. New York, NY: Springer; 2013, 151–66. [Google Scholar]
  • 26. Silva  AEP, Chiaravalloti Neto  F, Conceição  GMS.  Leptospirosis and its spatial and temporal relations with natural disasters in six municipalities of Santa Catarina, Brazil, from 2000 to 2016. Geospat Health. 2020;15:903. [DOI] [PubMed] [Google Scholar]
  • 27. Nardoni Marteli  A, Guasselli  LA, Diament  D  et al.  Spatio-temporal analysis of leptospirosis in Brazil and its relationship with flooding. Geospat Health. 2022;17:1128. [DOI] [PubMed] [Google Scholar]
  • 28. Teles  AJ, Bohm  BC, Silva  SCM  et al.  Spatial and temporal dynamics of leptospirosis in South Brazil: a forecasting and nonlinear regression analysis. PLoS Negl Trop Dis. 2023;17(4):e0011239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Phosri  A.  Effects of rainfall on human leptospirosis in Thailand: evidence of multi-province study using distributed lag non-linear model. Stoch Environ Res Risk Assess. 2022;36(12):4119–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. de Souza  KOC, Góes  JAP, Melo  MS  et al.  Spatiotemporal clustering, social inequities and the risk of leptospirosis in an endemic area of Brazil: a retrospective spatial modelling. Trans R Soc Trop Med Hyg. 2021;115(8):854–62. [DOI] [PubMed] [Google Scholar]
  • 31. Teles  AJ, Bohm  BC, Silva  SCM  et al.  Socio-geographical factors and vulnerability to leptospirosis in South Brazil. BMC Public Health. 2023;23(1):1311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Gutierrez  JD.  Effects of meteorological factors on human leptospirosis in Colombia. Int J Biometeorol. 2021;65(2):257–63. [DOI] [PubMed] [Google Scholar]
  • 33. Gutiérrez  JD, Martínez-Vega  RA.  Spatiotemporal dynamics of human leptospirosis and its relationship with rainfall anomalies in Colombia. Trans R Soc Trop Med Hyg. 2018;112(3):115–23. [DOI] [PubMed] [Google Scholar]
  • 34. Wang  RN, Zhang  YC, Yu  BT  et al.  Spatio-temporal evolution and trend prediction of the incidence of class B notifiable infectious diseases in China: a sample of statistical data from 2007 to 2020. BMC Public Health. 2022;22(1):1208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. López-Quílez  A.  Spatio-temporal analysis of infectious diseases. Int J Environ Res Public Health. 2019;16(4):669. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

trae013_Supplemental_File
trae013_supplemental_file.docx (1.7MB, docx)

Data Availability Statement

None.

Articles from Transactions of the Royal Society of Tropical Medicine and Hygiene are provided here courtesy of Oxford University Press

RESOURCES