Skip to main content
Download PDF

Relapses in canine leishmaniosis: risk factors identified through mixed-effects logistic regression

Parasites & Vectors volume 17, Article number: 357 (2024) Cite this article

Abstract

Background

Canine leishmaniosis (CanL), caused by Leishmania infantum, is an important vector-borne parasitic disease in dogs with implications for human health. Despite advancements, managing CanL remains challenging due to its complexity, especially in chronic, relapsing cases. Mathematical modeling has emerged as a powerful tool in various medical fields, but its application in understanding CanL relapses remains unexplored.

Methods

This retrospective study aimed to investigate risk factors associated with disease relapse in a cohort of dogs naturally infected with L. infantum. Data from 291 repeated measures of 54 dogs meeting the inclusion criteria were included. Two logistic mixed-effects models were created to identify clinicopathological variables associated with an increased risk of clinical relapses requiring a leishmanicidal treatment in CanL. A backward elimination approach was employed, starting with a full model comprising all potential predictors. Variables were iteratively eliminated on the basis of their impact on the model, considering both statistical significance and model complexity. All analyses were conducted using R software, primarily employing the lme4 package, and applying a significance level of 5% (P < 0.05).

Results

This study identified clinicopathological variables associated with an increased risk of relapses requiring a leishmanicidal treatment. Model 1 revealed that for each 0.1 increase in the albumin/globulin ratio (A/G) ratio, the odds of requiring treatment decreased by 45%. Conversely, for each unit increase in the total clinical score (CS), the odds of requiring treatment increase by 22–30%. Indirect immunofluorescence antibody test (IFAT) was not a significant risk factor in model 1. Model 2, incorporating individual albumin and globulins values, showed that dogs with high IFAT titers, hyper beta-globulinemia, hypoalbuminemia, anemia, and high CS were at increased risk of relapse. Both models demonstrated a good fit and explained a substantial amount of variability in treatment decisions.

Conclusions

Dogs exhibiting higher CS, dysproteinemia, anemia, and high IFAT titers are at increased risk of requiring leishmanicidal treatment upon clinical relapse in CanL. Regular monitoring and assessment of risk factors prove essential for early detection of relapses and effective intervention in CanL cases. The contrasting findings between the two models highlight the complexity of aspects influencing treatment decisions in this disease and the importance of tailored management strategies to improve outcomes for affected dogs.

Graphical Abstract

Background

Canine leishmaniosis (CanL), caused by Leishmania infantum, is among the most important vector-borne parasitic diseases of dogs, considered the main peridomestic reservoir of human infection for phlebotomine sand flies [1]. Despite advances in diagnosis and treatment, the management of CanL remains challenging due to the complexity of the disease, which often involves chronic, relapsing cases characterized by intermittent clinical manifestations and varying responses to therapy [2, 3]. Relapses, defined as the recurrence of clinical signs and/or clinicopathological abnormalities after an initial period of clinical improvement [4,5,6], represent a major concern in the clinical management of CanL. These relapses not only contribute to prolonged morbidity in sick dogs, but also represent challenges in terms of treatment efficacy and disease control [3]. Early identification of relapses is therefore crucial for optimizing therapeutic strategies and improving long-term outcomes in sick dogs.

In recent years, mathematical modelling approaches have emerged as powerful tools for predicting disease outcomes and guiding clinical decision-making in various medical fields, including infectious diseases [7]. By integrating epidemiological data, clinicopathological parameters, and host–parasite features, mathematical models offer the potential to identify predictive factors associated with infection and disease and to develop prognostic tools and individualized treatment strategies.

In the field of leishmaniosis, mathematical modeling has been used to identify risk factors for infection or seropositivity [8, 9], to explore genetic susceptibility to infection [10], to identify biomarkers for therapy outcomes [11], and to evaluate methods of control for human leishmaniasis [7], among others. However, to the best of our knowledge, no published studies have delved into the risk factors associated with disease relapse in CanL using mathematical models.

In this study, we aimed to identify potential risk factors of relapse among clinicopathological parameters commonly used in the monitoring of dogs with CanL, employing mixed-effects logistic regression modeling.

Methods

Inclusion criteria and selection of records

For this study, we assessed an electronic database of 1194 cases of canine leishmaniosis that sought consultation at the Veterinary Teaching Hospital of the Universidad Complutense de Madrid between 2010 and 2022 to identify those with the following inclusion criteria: (1) a confirmed diagnosis of L. infantum infection by indirect immunofluorescence antibody test (IFAT) or polymerase chain reaction (PCR) (IFAT ≥ 1:200 and/or PCR positive), (2) data on clinical signs and physical examination, (3) hematology and (4) plasma protein electrophoresis (PE) results, and (5) information about treatment regimen. Dogs affected by other vector-borne diseases (e.g., ehrlichiosis, anaplasmosis, babesiosis, dirofilariosis) were excluded to mitigate potential confounding factors, as the clinical signs and clinicopathological abnormalities associated with these conditions closely mimic those of leishmaniosis. Of those, 54 cases met the inclusion criteria and were selected for the study (Fig. 1).

Fig. 1
figure 1

Flow diagram of the inclusion procedure of the study

Full size image

Data extraction and processing

The medical records of the 54 dogs were analyzed to extract clinicopathological data obtained during each visit to our infectious diseases clinical unit. Raw data were stored in a table in.xlsx format. Demographic data, clinical signs, and physical examination findings were carefully recorded using a scoring system (Additional file 1: Table S1) ranging from 0 to 3 (indicating low to high severity) to obtain a total clinical score (CS) for each visit (maximum score of 74) (adapted from [12]).

Data on laboratory parameters were also extracted, including indirect immunofluorescence antibody test (IFAT) titers, hematocrit (%) (HCT), white blood cell count (× 103/\(\mu\) l) (WBC) and platelet count (× 103/\(\mu\) l) (PLT), total plasma proteins (g/dl), albumin (g/dl) (ALB), alpha-1 (g/dl), alpha-2 (g/dl), beta (g/dl), and gamma (g/dl) globulins, and albumin/globulin (A/G) ratio. Finally, information regarding the treatment regimen at each visit was also recorded. A relapse was defined as the initiation of a new cycle of treatment with meglumine antimoniate or miltefosine due to worsening of clinical signs and/or laboratory parameters [5, 6]. Dogs that did not receive a new treatment were considered clinically stable (no relapse).

Selection of variables

The outcome variable was binary, representing the receipt of treatment (no/yes) upon disease relapse. The clinicopathological parameters selected as risk factors associated with the likelihood of receiving treatment were selected on the basis of previous research [4, 12,13,14,15]. Numerical risk factors considered in the analysis included age, total proteins, ALB, alpha-1, alpha-2, beta, and gamma-globulins, A/G ratio, HCT, WBC, PLT, and CS. Categorical variables consisted of sex (male or female), breed (mongrel or purebred), and IFAT (negative: < 1:200, low positive: 1:200–1:400, medium positive: 1:800–1:1600, or high positive: ≥ 1:3200). IFAT cut-offs were defined according to values established by the laboratory [4, 16,17,18].

Statistical analysis and logistic regression model

For descriptive statistics, the median and interquartile ranges were calculated for all numeric data, given the non-normal distribution observed in the dataset. Additionally, frequency distributions were analyzed for categorical data.

Logistic mixed-effects model

The study employed a generalized linear mixed model (GLMM) framework to assess the impact of various predictors on the response variable treatment, assuming a binomial distribution. The GLMM accommodates repeated measurements on the same subject by incorporating a subject-specific random effect into the model, thereby capturing unobserved subject-specific characteristics. The logistic mixed-effects regression model is utilized to model binary outcome variables, wherein the log odds of the outcomes are expressed as a linear combination of the risk factors when both fixed and random effects are accounted for. The rationale for using repeated measures was the chronic feature of CanL and the fact that relapses can occur at any given time due to multiple reasons [19, 20].

Since regression models can be influenced by correlated variables, we tested for multicollinearity using the variance inflation factor (VIF) to ensure that explanatory variables were not highly correlated. Such a scenario would impede the accurate interpretation of the odds ratio (OR), as the outcome would no longer be solely influenced by an individual variable. All variables incorporated into the models demonstrated low VIF values (< 2) [21]. The dataset consisted of 291 observations nested within 54 subjects. Two models were developed:

Model 1 encompassed a full generalized mixed-effects model on treatment, aiming to incorporate one random effect (the intercepts of subjects) and nine fixed effects, namely, sex, breed, age, IFAT, A/G, HCT, WBC, PLT, and CS.

Model 2 encompassed a full generalized mixed-effects model based on treatment, attempting to integrate one random effect (the intercepts of subjects), and 13 fixed effects, namely, sex, breed, age, IFAT, ALB, alpha-1, alpha-2, beta, and gamma globulins, HCT, WBC, PLT, and CS.

Model fitting utilized maximum likelihood estimation with adaptive Gauss–Hermite quadrature. The aim of model selection was to identify significant predictors of the dependent variable while preventing overfitting. A backward elimination approach was employed, starting with a full model comprising all potential predictors. Variables were iteratively eliminated on the basis of their impact on the model, considering both statistical significance and model complexity. At each step, the variable with the highest P-value exceeding 0.15 and the least improvement in Akaike information criterion (AIC) was removed. AIC balances model complexity and goodness-of-fit, ensuring a trade-off between simplicity and predictive accuracy. The remaining variables were used to refine the model, and this process was repeated until all remaining variables had P-values below 0.15 and made meaningful contributions to the model according to the AIC [22]. After model fitting, an analysis of variance (ANOVA) was performed to compare the explained variance of each model generated during the backward elimination process. This provided a quantitative evaluation of how the model improved with each iteration and supported the selection of the final model.

The variance explained by the models was quantified as R2 marginal (indicating the variance explained by fixed effects) and R2 conditional (indicating the variance explained by both random and fixed effects), following a method tailored for generalized linear mixed-effects models [23]. All analyses were conducted using R software [24], employing the lme4 package [25], and applying a significance level of 5% (P < 0.05).

Results

Of the 54 dogs included in the study, 32 (59.2%) were male and 22 (40.7%) female, 18 (33.3%) were mongrel dogs and 37 (66.6%) were purebred, and their ages ranged from 1 to 13 years (median 6 years; IQR 4.30, 8.0). The number of visits for each dog ranged from 1 to 14 visits (median 4 visits; IQT 3, 7), resulting in a total of 291 visits. Follow-up recorded for the same dog ranged from 4.2 months to 9.4 years (median 1.96 years; IQR 0.9, 3.6 years).

Total clinical scores ranged between 0 and 18 (median 1; IQR 0, 3). While clinical signs were present at 151 visits (51.9%), dogs received leishmanicidal treatment at only 69 visits (23.7%). Medium IFAT titers (1:800–1:1600) were detected at most visits (128/291, 43.9%), followed by low positive titers (1:200–1:400) (99/291, 34%) and negative titers (< 1:200) (51/291, 18.5%). High antibody titers (≥ 1:3200) were detected in 13/291 visits (4.4%). The most common laboratory abnormality detected was hyper-gammaglobulinemia (49.8%), followed by decreased A/G ratio (38.1%), hyperproteinemia (33.3%), and leukopenia (33.3%) (Table 1).

Table 1 Distribution of laboratory parameters of the dogs included in the study
Full size table

Models

In this study, we aimed to investigate potential risk factors associated with clinical relapses requiring leishmanicidal treatment in CanL among a set of nine (model 1) and 13 (model 2) clinicopathological variables. A backward elimination approach was employed, starting with a full model comprising all potential predictors. Variables were iteratively eliminated on the basis of their impact on the model, considering both statistical significance and model complexity. At each step, the variable with the highest P-value exceeding 0.15 and the least improvement of AIC was removed. Given that the A/G ratio and individual globulins values are highly correlated variables, we created two separate models to avoid redundancy.

Model 1 retained only the A/G ratio and CS as fixed effects. In this model, the A/G ratio was inversely associated with the outcome. Notably, each increase of 0.1 in the A/G ratio was associated with a 45% decrease in the odds of receiving treatment due to a clinical relapse (OR 0.55, 95% CI 0.45–0.65, P < 0.001) (Table 2). Furthermore, the CS of the subjects was significantly associated with treatment (OR 1.22, 95% CI 1.08–1.37, P = 0.001), with a 22% increase in the odds of receiving treatment per each unit increase in the CS. Table 2 presents the odds ratios (OR) along with their corresponding 95% confidence intervals (CI) and P-values for both predictor variables.

Table 2 Odds ratio of the generalized linear mixed-effects found for model 1 on treatment with CS and A/G ratio included in the model
Full size table

Model fit was assessed using statistical measures, including marginal R2, conditional R2, and Akaike information criterion (AIC). Marginal R2 represents the variance explained by the fixed effects in the model and indicates how well the model fits the observed data. It is calculated as the proportion of variance explained by the fixed effects alone, without considering the random effects. In our study, model 1 explained 62% of the variance in the data (marginal R2), indicating a substantial amount of variability in the outcome variable (treatment receipt upon disease relapse) accounted by the predictor variables included in the model. Conditional R2 represents the variance explained by both the fixed and random effects in the model and provides a more comprehensive measure of model fit. It is calculated as the proportion of variance explained by both the fixed and random effects combined. In our study, model 1 explained 64% of the total variance when accounting for both fixed and random effects (conditional R2) (Table 2), indicating a substantial overall fit of the model to the data.

Model 2 retained IFAT, ALB, beta globulins, HCT, and CS as fixed effects. The second model revealed that dogs with high IFAT titers exhibit nearly 18 times greater odds (OR 17.98, 95% CI 1.64–197.28, P = 0.018) of requiring treatment compared with dogs with a negative IFAT (Table 3). Additionally, higher levels of beta globulins were associated with an increased chance of relapse (OR 1.10, 95% CI 1.03–1.17, P = 0.002), with every 0.1 increase in beta globulins increasing the odds of receiving treatment by 10%. Notably, the CS exhibited a significant positive association with the chance of treatment, with each unit increase in the CS associated with a 30% increase in the odds of receiving treatment (OR 1.30, 95% CI 1.15–1.48, P < 0.001). In contrast, higher levels of ALB (OR 0.89, 95% CI 0.82–0.96, P = 0.004) and HCT were associated with decreased odds of receiving treatment. (OR 0.93, 95% CI 0.88–0.99, P = 0.014). Specifically, for every 0.1 increase in albumin values, the odds of receiving treatment reduced by 11%, while for each unit (1.0) increase in HCT values they decreased by 7%. Lastly, model 2 explained 64% of the variance in the data (marginal R2) and 66% of the total variance when accounting for both fixed and random effects (conditional R2) (Table 3).

Table 3 Odds ratio of the generalized linear mixed effects found for model 2 on treatment, with globulins fractions included in the model, but not the A/G ratio
Full size table

Discussion

The results of the present study revealed important clinicopathological variables associated with an increased risk of clinical relapses requiring leishmanicidal treatment in dogs with CanL. The study cohort comprised 54 dogs, predominantly male purebreds with a median age of 6 years. The association of breed, sex, and age with infection or disease has been investigated in several studies, with conflicting results [26,27,28,29,30,31]. In our study, these variables were eliminated from both models in the backwards process, as they did not significantly improve the model’s predictive capacity in the presence of other more important variables.

Clinical signs were observed in more than half of the visits (51.9%), but a relapse was only present in one-third of them (23.7%). This is expected, as some clinical signs in CanL, such as uveitis and vasculitis, have chronic presentation and can take several weeks or even months to resolve. Nevertheless, the CS, derived from the sum of the presence and severity of various clinical signs, remained an integral part of both models, indicating that dogs with higher CS were more likely to require leishmanicidal treatment. Specifically, for every unit increase in the CS, the odds of requiring treatment increased from 22% to 30% (model 1 and model 2, respectively). This result underscores the potential value of employing a clinical scoring chart in the ongoing monitoring of dogs with CanL. While previous researchers have proposed different clinical scoring systems [12, 14, 32, 33], the absence of a validated scoring system for use in both clinical practice and research persists [18]. As pointed out by Meléndez-Lazo et al. (2018), an ideal scoring model should encompass relevant clinical signs and laboratory parameters, incorporating varying weights among parameters that could influence prognosis and survival, particularly those linked to renal function. Employing mathematical models could aid in the development of such scoring systems in future studies.

Regarding clinicopathological findings, the most common alterations detected were hyper-gamma globulinemia (49.8%), decreased A/G ratio (38.1%), hyperproteinemia (33.3%), and leukopenia (33.3%), in agreement with previous research [15, 18, 34, 35]. In contrast, alpha‐1 and alpha-2 globulins were elevated in only a small number of visits (3.8% and 3.1%, respectively) (Table 1). One study found a significant increase in alpha‐2 globulins in dogs with leishmaniosis, however, dogs were not treated and were followed for 1 month only [13]. Paltrinieri et al. (2016) also described a moderate increase in alpha-2 globulins in early stages of the disease. The stage of the disease at which the dogs were assessed in each study and/or differences in PE techniques used could explain this discrepancy.

In the first model, the A/G ratio emerged as a significant predictor of relapse. Notably, each increase of 0.1 in the A/G ratio was associated with a 45% decrease in the odds of receiving treatment due to a clinical relapse. Protein electrophoresis has been proven to be extremely useful for the diagnosis and monitoring of CanL and is routinely performed in clinical practice [15, 32, 36, 37]. In fact, previous studies demonstrated that this technique may show abnormalities very early during the course of the disease and even before the onset of overt clinical signs [13, 32, 35, 36, 38, 39].

Interestingly, IFAT titers were not statistically significant, and therefore were not retained in model 1. This indicates that in the presence of a low A/G ratio and clinical signs suggestive of CanL, the IFAT titer may not be a robust predictor of relapse. Even though previous research has established an association between high antibody titers and clinical disease [2, 15, 40], practitioners must be aware of IFAT’s limitations when using it for monitoring treatment efficacy or detecting clinical relapses. For instance, serology may not be a useful parameter in the short term as antibody titers may take up to 6 months to substantially decrease [12, 15, 38, 41]. Additionally, the dynamics of IFAT titers vary on the basis of the dog’s inherent response to treatment. Dogs with a favorable response typically exhibit a reduction in antibody titers [38, 42, 43], whereas those with suboptimal responses may maintain chronically elevated titers or experience only minor or temporary decreases [4, 44,45,46]. Furthermore, an increase in antibody levels may occur in the presence of other infections or non-infectious diseases such as endocrinopathies or neoplasia that trigger the multiplication of Leishmania parasites [47]. Consequently, practitioners should not rely solely on an elevation in IFAT titers to initiate leishmanicidal treatment, as evidenced by the performance of our models incorporating other variables.

The substitution of the A/G ratio with globulins fractions in model 2 led to interesting results. In the absence of this variable, the presence of high antibody titers (> 1:1600) was significantly associated with relapse. Low albumin, considered a negative prognostic factor in CanL [48], was a common finding (22.7%) in our study, and an increase in its values was associated with a decreased risk of relapse. In contrast, hyperbetaglobulinemia was associated with an increased risk, despite its relatively low frequency in the study (6.2%). Surprisingly, gamma-globulins were not retained in model 2, despite hypergammaglobulinemia being the most common laboratory disorder detected (49.8%). One explanation for this finding is the moderate correlation between gamma-globulins and HCT (r = −0.53) (data not shown), the last being retained in the model. When two variables are highly correlated, the model may have difficulty distinguishing the unique contributions of each variable to the outcome and may choose to retain only one of them to avoid multicollinearity issues. Therefore, considering the possibility that most anemic dogs also had hypergammaglobulinemia, the last was removed in the stepwise process for not significantly contributing to the model’s predictive capability in the presence of the other variables.

In model 2, HCT also emerged as a significant predictor of relapse. Anemia, present in 15.1% of visits in this study, is one of the most frequent clinicopathological abnormalities detected in CanL and is often associated with the presence of clinical signs [13, 49]. However, in the first model, HCT did not significantly enhance the model’s predictive capacity, indicating that in the presence of a low A/G ratio and clinical signs, the presence of anemia is a less critical factor to consider before initiating leishmanicidal treatment in suspected relapse cases.

The decision to develop two separate models, one including the A/G ratio and the other incorporating individual albumin and globulins values, was driven by the high collinearity between these variables. Furthermore, considering that determining all protein fractions may not always be feasible due to equipment availability or cost constraints, opting for two models seemed plausible. The contrasting findings between the two models highlight the need for a comprehensive approach to risk assessment in CanL clinical management.

The assessment of model fit using statistical measures such as the marginal R2, conditional R2, and AIC provided insights into the performance of the models. The relatively high marginal R2 and conditional R2 values indicate that the models explained a substantial amount of variability in the outcome variable, accounting for both fixed and random effects.

It is important to recognize the limitations of our study, including the relatively small sample size and the retrospective nature of the data analysis. Future studies with larger sample sizes and prospective designs are warranted to validate our findings and enhance our understanding of the factors influencing relapses in CanL. Additionally, the characteristics of the study population being limited to dogs referred to a specialist in infectious and parasitic diseases may affect the model’s performance in other populations of dogs [50]. Lastly, the presence of other diseases can potentially influence the performance of our model, and this should be taken into consideration in future studies. Therefore, in suspected relapses, the presence of other infectious, metabolic, or neoplastic diseases that could impair the immune system should always be considered before starting a leishmanicidal treatment, even in the presence of the clinicopathological alterations considered significant in this study. Focusing on CanL as an isolated disease may lead to poor clinical management and treatment response.

Early detection of relapses in CanL is crucial, ensuring timely treatment initiation, thereby preventing disease progression and improving outcomes [17]. Prompt intervention may not only reduce the parasite burden [51], lowering the risk of transmission, but also mitigate complications, enhancing the quality of life for infected dogs. Therefore, regular monitoring and assessment of risk factors prove essential for early detection of relapses and effective intervention in CanL clinical cases.

The findings of this study contribute to our understanding of the risk factors associated with clinical relapses requiring leishmanicidal treatment in CanL. By elucidating the roles of clinical scores, plasma protein components, HCT, and IFAT titers through model development, our study provides valuable insights for clinicians and researchers aiming to optimize treatment strategies and improve outcomes in CanL clinical management and decision-making.

Conclusions

This study reveals important clinicopathological variables associated with an increased risk of clinical relapses requiring leishmanicidal treatment in CanL-sick dogs, providing valuable insights for clinicians and researchers. Dogs exhibiting higher total clinical scores, low A/G ratio, hypoalbuminemia, hyperbetaglobulinemia, anemia, and high IFAT titers are at increased risk of requiring leishmanicidal treatment due to a clinical relapse, underscoring the importance of monitoring these parameters in dogs with CanL. The contrasting findings between the two models highlight the complexity of factors influencing treatment decisions in this disease, emphasizing the need for tailored clinical management strategies to improve outcomes for sick dogs. Future studies with larger sample sizes and prospective designs are warranted to validate our findings and enhance our understanding of the factors influencing relapses in CanL.

Availability of data and materials

The raw data and R script with the specific codes used in this study will be available upon request.

Abbreviations

CanL:

Canine leishmaniosis

IFAT:

Indirect immunofluorescence antibody test

PCR:

Polymerase chain reaction

PE:

Protein electrophoresis

A/G ratio:

Albumin/globulin ratio

GLMM:

Generalized linear mixed-model

OR:

Odds ratio

CI:

Confidence interval

AIC:

Akaike information criterion

VIF:

Variance inflation factor

HCT:

Hematocrit

WBC:

White blood cell count

PLT:

Platelet count

CS:

Clinical score

R2:

Coefficient of determination

References

  1. Otranto D, Dantas-Torres F. The prevention of canine leishmaniasis and its impact on public health. Trends Parasit. 2013;29:339–45.

    Article  Google Scholar 

  2. Manna L, Corso R, Galiero G, Cerrone A, Muzj P, Gravino AE. Long-term follow-up of dogs with leishmaniosis treated with meglumine antimoniate plus allopurinol versus miltefosine plus allopurinol. Parasit Vectors. 2015;8:289.

    Article  PubMed  PubMed Central  Google Scholar 

  3. Yasur-Landau D, Jaffe CL, David L, Baneth G. Allopurinol resistance in Leishmania infantum from dogs with disease relapse. PLoS Negl Trop Dis. 2016;10:e0004341.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Solano-Gallego L, Miró G, Koutinas A, Cardoso L, Pennisi MG, Ferrer L, Bourdeau P, Oliva G, Baneth G, The LeishVet Group. LeishVet guidelines for the practical management of canine leishmaniosis. Parasit Vectors. 2011;4:86.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Cota GF, de Sousa MR, Rabello A. Predictors of visceral leishmaniasis relapse in HIV-infected patients: a systematic review. PLoS Negl Trop Dis. 2011;5:e1153.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Iarussi F, Paradies P, Foglia Manzillo V, Gizzarelli M, Caratozzolo MF, Navarro C, Greco B, Rubino GTR, Oliva G, Sasanelli M . Comparison of two dosing regimens of miltefosine, both in combination with allopurinol, on clinical and parasitological findings of dogs with leishmaniosis: a pilot study. Front Vet Sci. 2020;7:577–395.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Rock KS, Quinnell RJ, Medley GF, Courtenay O. Chapter two - Progress in the mathematical modelling of visceral leishmaniasis. Adv Parasitol. 2016;94:49–131.

    Article  CAS  PubMed  Google Scholar 

  8. Coura-Vital W, Reis AB, Reis LES, Braga SL, Roatt BM, de Aguiar-Soares RDO, Marques MJ, Veloso VM, Carneiro M. Canine visceral leishmaniasis: incidence and risk factors for infection in a cohort study in Brazil. Vet Parasitol. 2013;197:411–7.

    Article  PubMed  Google Scholar 

  9. Gouvêa MV, Mendonça IL, do Maria Cruz SPE, Costa CHN, Braga JU, Werneck GL. Predictive factors for Leishmania infantum infection in dogs examined at a veterinary teaching hospital in Teresina State of Piauí Brazil. Rev Soc Bras Med Trop. 2016;49:107–11.

    Article  PubMed  Google Scholar 

  10. Quilez J, Martínez V, Woolliams JA, Sanchez A, Pong-Wong R, Kennedy LJ, Quinnell RJ, Ollier WER, Roura X, Ferrer L, Altet L, Francino O. Genetic control of canine leishmaniasis: genome-wide association study and genomic selection analysis. PLoS ONE. 2012;7:e35349.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. de Sousa GR, Alves de Pinho F, Dinis-Oliveira RJ, Azevedo R, Gaifem J, Farias Larangeira D, Ramos-Sanchez EM, Goto H, Silvestre R, Barrouin-Melo, SM. Mathematical modelling using predictive biomarkers for the outcome of canine leishmaniasis upon chemotherapy. Microorganisms. 2020;8:745.

    Article  Google Scholar 

  12. Miró G, Oliva G, Cruz I, Cañavate C, Mortarino M, Vischer C, Bianciardi P. Multicentric, controlled clinical study to evaluate effectiveness and safety of miltefosine and allopurinol for canine leishmaniosis. Vet Dermatol. 2009;20:397–404.

    Article  PubMed  Google Scholar 

  13. Ribeiro RR, da Silva SM, de Fulgêncio GO, Michalick MSM, Frézard FJG. Relationship between clinical and pathological signs and severity of canine leishmaniasis. Rev Bras Parasitol Vet. 2013;22:373–8.

    Article  PubMed  Google Scholar 

  14. Proverbio D, Spada E, Bagnagatti de Giorgi G, Perego R, Valena E. Relationship between Leishmania IFAT titer and clinicopathological manifestations (clinical score) in dogs. Biomed Res Int. 2014;2014:412808.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Paltrinieri S, Gradoni L, Roura X, Zatelli A, Zini E. Laboratory tests for diagnosing and monitoring canine leishmaniasis. Vet Clin Pathol. 2016;45:552–78.

    Article  PubMed  Google Scholar 

  16. Mancianti F, Meciani N. Specific serodiagnosis of canine leishmaniasis by indirect immunofluorescence, indirect hemagglutination, and counterimmunoelectrophoresis. Am J Vet Res. 1988;49:1409–11.

    CAS  PubMed  Google Scholar 

  17. Solano-Gallego L, Koutinas A, Miró G, Cardoso L, Pennisi MG, Ferrer L, Bordeau P, Oliva G, Baneth G. Directions for the diagnosis, clinical staging, treatment and prevention of canine leishmaniosis. Vet Parasitol. 2009;165:1–18.

    Article  CAS  PubMed  Google Scholar 

  18. Meléndez-Lazo A, Ordeix L, Planellas M, Pastor J, Solano-Gallego L. Clinicopathological findings in sick dogs naturally infected with Leishmania infantum: Comparison of five different clinical classification systems. Res Vet Sci. 2018;117:18–27.

    Article  PubMed  Google Scholar 

  19. Baneth G, Shaw SE. Chemotherapy of canine leishmaniosis. Vet Parasitol. 2002;106:315–24.

    Article  CAS  PubMed  Google Scholar 

  20. Miró G, Cardoso L, Pennisi MG, Oliva G, Baneth G. Canine leishmaniosis—New concepts and insights on an expanding zoonosis: part two. Trends Parasitol. 2008;24:371–7.

    Article  PubMed  Google Scholar 

  21. Johnston R, Jones K, Manley D. Confounding and collinearity in regression analysis: a cautionary tale and an alternative procedure, illustrated by studies of British voting behaviour. Qual Quant. 2018;52:1957–76.

    Article  PubMed  Google Scholar 

  22. Bursac Z, Gauss CH, Williams DK, Hosmer DW. Purposeful selection of variables in logistic regression. Source Code Biol Med. 2008;3:17.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Nakagawa S, Schielzeth H. A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods in Ecol and Evol. 2013;4:133–42.

    Article  Google Scholar 

  24. R Core Team. (2024) The R Project for Statistical Computing. R: The R Project for Statistical Computing. https://www.r-project.org. Accessed 20 Mar 2024.

  25. Bates D, Maechler M, Bolker B, Walker S. Linear mixed-effects models using Eigen and S4. R package version 1.0–4. 2023. https://cran.r-project.org/web/packages/lme4/lme4.pdf. Accessed 7 Feb 2024.

  26. Miranda S, Roura X, Picado A, Ferrer L, Ramis A. Characterization of sex, age, and breed for a population of canine leishmaniosis diseased dogs. Res Vet Sci. 2008;85:35–8.

    Article  PubMed  Google Scholar 

  27. Gálvez R, Descalzo MA, Miró G, Jiménez MI, Martín O, Dos Santos-Brandao F, Guerrero I, Cubro E, Molina R. Seasonal trends and spatial relations between environmental/meteorological factors and leishmaniosis sand fly vector abundances in Central Spain. Acta Trop. 2010;115:95–102.

    Article  PubMed  Google Scholar 

  28. Cabré M, Planellas M, Ordeix L, Solano-Gallego L. Is signalment associated with clinicopathological findings in dogs with leishmaniosis? Vet Rec. 2021;189:e451.

    Article  PubMed  Google Scholar 

  29. Silva SS, de Macedo LO, de Oliveira JCP, Alves LC, de Carvalho GA, Ramos RAN. Canine visceral leishmaniasis: risk factors and spatial analysis in an endemic area of Northeastern Brazil. Rev Bras Parasitol Vet. 2023;32:e003223.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Dos Reis AT, do Carmo Silva CF, Rocha TB, Coutinho DJB, da Costa AP, Costa FB, Souza FA, Nogueira RMS. Exposure to and infection by Leishmania infantum among domestic dogs in an area of the Cerrado biome, Maranhão Brazil. Vet Parasit Reg Stud Rep. 2023;39:100851.

    Google Scholar 

  31. Chiyo L, Dos Santos AG, de Souza AB, Rivas AV, Valle SB, da Sevá AP, Viana KF. Cross-sectional spatial and epidemiological analysis of canine visceral leishmaniasis cases in the triple border region, Brazil, Argentina and Paraguay, between 2015 and 2020. Acta Trop. 2023;239:106811.

    Article  PubMed  Google Scholar 

  32. Paradies P, Sasanelli M, de Caprariis D, Testini G, Traversa D, Lia RP, Danta-Torres F, Otranto D. Clinical and laboratory monitoring of dogs naturally infected by Leishmania infantum. Vet J. 2010;186:370–3.

    Article  PubMed  Google Scholar 

  33. da Silva KR, de Mendonça VRR, Silva KM, do Nascimento LFM, Mendes-Sousa AF, de Pinho FA, Barral-Netto M, Barral AMP, Pires e Cruz MS. Scoring clinical signs can help diagnose canine visceral leishmaniasis in a highly endemic area in Brazil. Mem Inst Oswaldo Cruz. 2017;112:53–63.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Ciaramella P, Oliva G, De Luna R, Ambrosio R, Cortese L, Persechino A, Gradoni L, Scalone A. A retrospective clinical study of canine leishmaniasis in 150 dogs naturally infected by Leishmania infantum. Vet Record. 1997;141:539–43.

    Article  CAS  PubMed  Google Scholar 

  35. Almeida MAO, Jesus EEV, Sousa-Atta MLB, Alves LC, Berne MEA, Atta AM. Clinical and serological aspects of visceral leishmaniasis in northeast Brazilian dogs naturally infected with Leishmania chagasi. Vet Parasit. 2005;127:227–32.

    Article  CAS  Google Scholar 

  36. Riera C, Valladares JE, Gállego M, Aisa MJ, Castillejo S, Fisa R, Ribas N, Carrió J, Alberola J, Arboix M. Serological and parasitological follow-up in dogs experimentally infected with Leishmania infantum and treated with meglumine antimoniate. Vet Parasit. 1999;84:33–47.

    Article  CAS  Google Scholar 

  37. Villanueva-Saz S, Ballesté C, Casasnovas AF, Estevez AY, Arribas MTV. Bisalbuminemia in a dog with leishmaniosis after anti-Leishmania therapeutic protocol administration: a rare condition detected in the electrophoretogram. Vet Record Case Rep. 2020;8:e001290.

    Article  Google Scholar 

  38. Torres M, Bardagí M, Roura X, Zanna G, Ravera I, Ferrer L. Long term follow-up of dogs diagnosed with leishmaniosis (clinical stage II) and treated with meglumine antimoniate and allopurinol. Vet J. 2011;188:346–51.

    Article  CAS  PubMed  Google Scholar 

  39. Foglia Manzillo V, Di Muccio T, Cappiello S, Scalone A, Paparcone R, Fiorentino E, Gizzarelli M, Gramiccia M, Gradoni L, Oliva G. Prospective study on the incidence and progression of clinical signs in naïve dogs naturally infected by Leishmania infantum. PLoS Negl Trop Dis. 2013;7:e2225.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Oliva G, Scalone A, Foglia Manzillo V, Gramiccia M, Pagano A, Di Muccio T, Gradoni L. Incidence and time course of Leishmania infantum infections examined by parasitological, serologic, and nested-PCR techniques in a cohort of naïve dogs exposed to three consecutive transmission seasons. J Clin Microbiol. 2006;44:1318–22.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Ferrer L, Aisa MJ, Roura X, Portús M. Serological diagnosis and treatment of canine leishmaniasis. Vet Rec. 1995;136:514–6.

    Article  CAS  PubMed  Google Scholar 

  42. Pasa S, Toz SO, Voyvoda H, Ozbel Y. Clinical and serological follow-up in dogs with visceral leishmaniosis treated with allopurinol and sodium stibogluconate. Vet Parasitol. 2005;128:243–9.

    Article  CAS  PubMed  Google Scholar 

  43. Cantos-Barreda A, Escribano D, Martínez-Subiela S, Pardo-Marín L, Segarra S, Cerón JJ. Changes in serum anti-Leishmania antibody concentrations measured by time-resolved immunofluorometric assays in dogs with leishmaniosis after treatment. Vet Immunol Immunopathol. 2018;198:65–9.

    Article  CAS  PubMed  Google Scholar 

  44. Oliva G, Gradoni L, Cortese L, Orsini S, Ciaramella P, Scalone A, de Luna R, Persechino A. Comparative efficacy of meglumine antimoniate and aminosidine sulphate, alone or in combination, in canine leishmaniasis. Ann Trop Med Parasitol. 1998;92:165–71.

    Article  CAS  PubMed  Google Scholar 

  45. Solano-Gallego L, Riera C, Roura X, Iniesta L, Gallego M, Valladares JE, Fisa R, Castillejo S, Alberola J, Ferrer L, Arboix M, Portús M. Leishmania infantum-specific IgG, IgG1 and IgG2 antibody responses in healthy and ill dogs from endemic areas: evolution in the course of infection and after treatment. Vet Parasitol. 2001;96:265–76.

    Article  CAS  PubMed  Google Scholar 

  46. Cantos-Barreda A, Escribano D, Egui A, Thomas MC, López MC, Tecles F, Bernal LJ, Cerón JJ, Martínez-Subiela S. One-year follow-up of anti-Leishmania antibody concentrations in serum and saliva from experimentally infected dogs. Int J Parasit. 2019;49:893–900.

    Article  CAS  Google Scholar 

  47. Foglia Manzillo V, Pagano A, Guglielmino R, Gradoni L, Restucci B, Oliva G. Extranodal γδ-T-cell lymphoma in a dog with leishmaniasis. Vet Clin Pathol. 2008;37:298–301.

    Article  PubMed  Google Scholar 

  48. Geisweid K, Mueller R, Sauter-Louis C, Hartmann K. Prognostic analytes in dogs with Leishmania infantum infection living in a non-endemic area. Vet Rec. 2012;171:399.

    Article  CAS  PubMed  Google Scholar 

  49. de Freitas JCC, Nunes-Pinheiro DCS, Lopes Neto BE, Santos GJL, de Abreu CRA, Braga RR, Campos Rde M, Oliveira LF. Clinical and laboratory alterations in dogs naturally infected by Leishmania chagasi. Rev Soc Bras Med Trop. 2012;45:24–9.

    Article  PubMed  Google Scholar 

  50. Stoltzfus JC. Logistic regression: a brief primer. Acad Emerg Med. 2011;18:1099–104.

    Article  PubMed  Google Scholar 

  51. Miró G, Gálvez R, Fraile C, Descalzo MA, Molina R. Infectivity to Phlebotomus perniciosus of dogs naturally parasitized with Leishmania infantum after different treatments. Parasit Vectors. 2011;4:52.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgement

The authors would like to thank the entire staff of the Complutense Veterinary Teaching Hospital, Spain, especially Pilar Sagredo, Cristina Fraile, Cristina Rupérez, Ana Montoya, Araceli Fernández, and Rosario Sanz from the Clinical Unit of Infectious Diseases, for their contribution to clinical data collection.

Funding

Open access funding provided by ELANCO. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author information

Authors and Affiliations

  1. Animal Health Department, Faculty of Veterinary Medicine, Complutense University of Madrid, Madrid, Spain

    Juliana Sarquis, Carolina R. Sanz, Ana Montoya, Juan Pedro Barrera, Rocío Checa & Guadalupe Miró

  2. Department of Quantitative Methods, Universidade Federal Do Estado Do Rio de Janeiro, Rio de Janeiro, Brazil

    Letícia Martins Raposo

  3. Departament of Medicine and Animal Surgery, Faculty of Veterinary Medicine, Complutense University of Madrid, Madrid, Spain

    Blanca Perez-Montero & María Luisa Fermín Rodríguez

Authors
  1. Juliana Sarquis
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Letícia Martins Raposo
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Carolina R. Sanz
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Ana Montoya
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Juan Pedro Barrera
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Rocío Checa
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Blanca Perez-Montero
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. María Luisa Fermín Rodríguez
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Guadalupe Miró
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

J.S. assisted with the clinical cases, performed data extraction, cleaning and curation, formal analysis, investigation, writing, review, and editing. G.M. contributed to methodology development and review and editing of the original draft. L.M.R. provided formal analysis, including model development, and contributed to the review and editing of the original draft. C.R.S. participated in data extraction, methodology development, and review and editing of the original draft. G.M., A.M., J.P.B., R.C., B.P.M., and M.L.F.R. contributed to the review and editing of the original draft. B.P.M. and M.L.F.R. also conducted laboratory analysis of patient samples that originated the data of this study. G.M. is the chief of the Infectious Diseases Unit, overseeing patient care and personally contributing to data generation for this retrospective study. G.M. proposed and designed the study, managed the clinical cases, and also provided supervision, project administration, and funding acquisition for publication.

Corresponding author

Correspondence to Guadalupe Miró.

Ethics declarations

Ethics approval and consent to participate

Not applicable. No live animals were involved in this research. We utilized existing data from medical records and laboratory reports, focusing exclusively on the analysis of anonymized data. Permission to access the data was obtained from the chief of service at the hospital where the consultations took place.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Sarquis, J., Raposo, L.M., Sanz, C.R. et al. Relapses in canine leishmaniosis: risk factors identified through mixed-effects logistic regression. Parasites Vectors 17, 357 (2024). https://doi.org/10.1186/s13071-024-06423-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s13071-024-06423-1

Keywords

Parasites & Vectors

ISSN: 1756-3305

Contact us