Synovial fluid and radiographic evaluation of joints from dogs with visceral leishmaniasis

Polyarthritis has been associated with canine visceral leishmaniasis (CanVL), and co-infection with Ehrlichia canis is common and may alter clinical manifestations. A total of 89 dogs presenting CanVL were subdivided into two groups: (1) G1, consisting of 46 dogs seronegative to Ehrlichia spp., and (ii) G2, consisting of 43 dogs seropositive to Ehrlichia spp. Eight joints (carpal, tarsal, stifles and elbows) from each dog were evaluated by radiography and synovial fluid (SF) cytologic analysis. Overall, 74 of the 89 (83.1%) dogs presented joint abnormalities suggestive of osteoarthritis by radiography (G1: 40/46 [86.9%]; G2: 34/43 [79.0%]), with no statistically significant between-group difference. All dogs with abnormal joint X-ray images presented radiographic lesions bilaterally, independent of the characteristics of the lesion. Soft tissue swelling around the joint and joint space narrowing were more commonly observed in G1 than in G2 dogs. There was no significant between-group difference in terms of other radiographic abnormalities suggestive of osteoarthritis (evident trabecular pattern, subchondral bone sclerosis, osteolysis, osteolytic–proliferative lesions or bone proliferation). SF from 174/315 (55.2%) and 152/307 (49.5%) joints from G1 and G2 dogs, respectively, presented an inflammatory infiltrate, but there was no significant association between the presence of inflammatory infiltrate and group. There was also no statistical difference between groups in either of the evaluated joints in terms of the percentage of neutrophils or mononuclear cells. Leishmania spp. amastigotes were found in 69/315 (21.9%) joints from G1 dogs and in 100/307 (32.5%) joints from G2 dogs (Fisherʼs exact test, P = 0.002, odds ratio = 0.5, 95% confidence interval = 0.4–0.8). The neutrophilic infiltrate was significantly higher in joints with amastigote forms in both G1 (Mann–Whitney U-test, U(18) = 817, Z = -3.76, P = 0.0001) and G2 dogs (Mann–Whitney U-test, U(18) = 6543, Z = − 5.06, P < 0.0001). A high prevalence of arthritis in dogs with CanVL was found, and all dogs presented involvement in multiple joints. Although no difference was observed between groups in terms of the number of dogs with polyarthritis and the presence of an inflammatory infiltrate in SF, Leishmania spp. amastigotes were found more frequently in joints from G2 dogs. Further studies evaluating SF in dogs co-infected with L. infantum and E. canis should be performed to evaluate this finding.


Background
Canine visceral leishmaniasis (CanVL) is a chronic and progressive disease caused by Leishmania infantum. Its presentation is variable and can include such clinical signs as fever, progressive weight loss, peripheral lymphadenopathy, splenomegaly, dermatological lesions,
Arthritis and polyarthritis have been associated with CanVL and may be caused by two mechanisms, the direct presence of L. infantum amastigotes and/or a type III hypersensitivity reaction with deposition of immune complex within the joint [4,5]. These two mechanisms are probably interrelated and may coexist in the same joint [6]. Leishmania infantum can cause either non-erosive [4,[7][8][9] or erosive [4,8,[10][11][12] arthritis. While the presence of bone and joint lesions due to CanVL in plain radiographs has been reported, computed tomography findings have suggested that joint lesions may be common in CanVL but not so commonly described because of the subclinical characteristic observed in most cases [13].
Ehrlichia canis-related polyarthritis in dogs with detection of Ehrlichia morulae in the synovial fluid (SF) has been reported in dogs, but only rarely [14,15], and there is no conclusive evidence supporting the occurrence of arthritis [16]. In addition, SF evaluation revealed no cytological evidence of arthritis in dogs with experimental acute canine monocytic ehrlichiosis (CME) [17]. Accordingly, the aim of this study was to evaluate the joint involvement pattern, by radiography and SF cytologic examination, of dogs presenting CanVL with and without the presence of anti-Ehrlichia spp. antibodies.

Animal selection and CanVL diagnosis
A total of 89 adult dogs naturally affected and presenting clinical signs of CanVL that had been referred to the Center for Zoonosis Control of Araçatuba (21°12′32″ S and 50°25′58″ W), a municipality with high endemicity for CanVL in São Paulo State, southern Brazil, were evaluated. CanVL diagnosis was based on the direct observation of Leishmania amastigote forms in smears from lymph node aspirates obtained by fine-needle aspiration and stained with a quick Romanovsky-type stain (Panótico Rápido ® ; Laborclin, Pinhais, Brazil), and on serology by using indirect enzyme-linked immunosorbent assay (ELISA) kit (Bio-Manguinhos/Fiocruz, Rio de Janeiro, Brazil). The ELISA is based on microtiter plates previously sensitized with soluble antigens of Leishmania major (e.g. MHOM/BR/76/JOF) and is used by the Brazilian Ministry of Health for the diagnosis of CanVL. The inclusion criterion for the study was positive test results for CanVL in both diagnostic tests. In compliance with the Brazilian federal law at the time of the study, dogs were euthanized after L. infantum diagnosis. All dogs underwent radiographic examination and arthrocentesis of eight different joints before euthanasia.
All dogs were tested for the presence of anti-Ehrlichia spp., anti-Anaplasma spp. and anti-Borrelia burgdorferi antibodies by a commercial ELISA rapid test (SNAP ® 3Dx ® ; IDEXX Laboratories Inc., Westbrook, ME, USA), according to the manufacturer's instructions. All serum samples tested negative for anti-Anaplasma spp. and anti-B. burgdorferi antibodies. According to the presence or absence of anti-Ehrlichia spp. antibodies, dogs were subdivided into two groups: (i) Group 1 (G1), which included 46 dogs with CanVL and seronegative to Ehrlichia spp.; and (ii) Group 2 (G2), which included 43 dogs with CanVL and seropositive to Ehrlichia spp.

Imaging studies
In both the right and left limbs, craniocaudal and mediolateral projections of the elbow and stifle joints and dorsopalmar/dorsoplantar and mediolateral projections of the carpal and tarsal joints were performed with an X-ray system (SHF 730; CRX, São Paulo, Brazil), with total imaging of eight joints per dog. The parameters were a focus-film distance of 10 cm; 10 mAs and variable kilovoltage peak (kVp; range: 50-70) according to the measured thickness of each joint region. Films were automatically processed using an MX2 processor (Macrotec, São Paulo, Brazil). Each joint was examined for the presence of lesions, type of lesions and involvement of adjacent soft tissues. Two experienced veterinary radiologists evaluated all images.

SF examination
Aseptic arthrocentesis of elbows, carpals, stifles and tarsal joints were performed using a fine needle (25 × 0.7 mm) and a 5-ml sterile syringe, and the maximum amount of SF was collected. SF was obtained by arthrocentesis of at least four joints of each dog. SF samples from the remaining joints were discarded due to insufficient volume or the presence of blood in the sample. The slides were prepared immediately after SF collection by cytocentrifugation at 380 g for 5 min. Direct cytologic examination was performed on smears stained by a quick Romanovskytype stain (Panótico Rápido ® ; Laborclin). Ten fields were examined in each smear, and the cell count was recorded. The percentage of nucleated cells (neutrophils, lymphocytes, monocytes and macrophages) was calculated, and the presence of Leishmania spp. amastigotes was determined under a light microscope at ×1000 magnification. The SF was considered to be normal if no more than 3 cells/field at ×1000 magnification were noted [18].

Statistical analysis
Descriptive data were tabulated in an electronic spreadsheet (Excel ® ; Microsoft Corp., Redmond, WA, USA), with absolute and relative (%) frequencies of radiographic and SF findings determined. Comparisons between groups were performed after analyzing the variables for normality (Shapiro-Wilk test) using the Mann-Whitney test. Associations between categorical variables were determined by Fisher's exact test with the determination of the odds ratio (OR) and 95% confidence interval (CI). Analyses were performed using commercially available statistical software (InStat version 6.00; GraphPad Software, San Diego, CA, USA) and considered significant when P < 0.05.

Animals and groups
The dogs included in this study had been referred to the Municipal Zoonoses Control Center due to a diagnosis of CanVL at a time when euthanasia of these animals was recommended in Brazil; consequently, it was not possible to obtain their clinical history. However, all dogs presented multiple clinical signs of CanVL at the time of sampling, such as weight loss, skin lesions, onychogryphosis, epistaxis, lymphadenopathy and diarrhea.

SF examination
The SF was collected without complications (sufficient volume and absence of blood in the sample) from 315 of 368 (85.5%) joints from G1 dogs and from 307 of 344 (89.2%) joints from G2 dogs. SF collected from 174 of 315 (55.2%; 95% CI = 49.7-60.6) joints from G1 dogs and from 152 of 307 (49.5%; 95% CI = 43.9-55.0) joints from G2 dogs presented an inflammatory infiltrate without a significant association between the presence of inflammatory infiltrate and the groups (neutrophils: Mann-Whitney U-test, U (18)  All joints that presented SF with an inflammatory infiltrate also had radiographic evidence of osteoarthritis. Many macrophages contained magenta-staining organisms, 2-3 µm in length with a kinetoplast, compatible with Leishmania spp. amastigotes, were also visualized (Fig. 2). Leishmania spp. amastigotes were observed free and within macrophages.  (Table 1). Per dog, the range of joints infected with Leishmania spp. amastigotes varied from one to eight in both groups, and the percentage of infected joints among the evaluated joints was similar (tarsal: 28.6%; carpal: 28.0%; elbow: 26.8%; stifle: 24.8%). The neutrophilic infiltrate was significantly higher in joints with Leishmania spp. amastigote forms in both G1 (Mann-Whitney U-test, U (18) = 8174, Z = -3.76,  (Fig. 3).

Discussion
Immune-mediated polyarthritis (IMPA) is the most common form of inflammatory joint disease found in dogs. Differential diagnosis of secondary IMPA includes vector-borne diseases (specifically those caused by B. burgdorferi and L. infantum) and sulfonamide-containing medications [19]. Additionally, E. canis-related polyarthritis has been rarely reported in dogs [14,15], although no previous studies have consistently supported the occurrence of this clinical sign in CME [16,17]. In Brazil, E. canis is the most common and widespread tick-borne pathogen in dogs [20]. Although cases of borreliosis-like disease have been reported in humans and animals by western blotting and enzyme immunoassay tests [21], to date Borrelia spp. has never been detected by molecular methods or isolated by cell culture in dogs in Brazil, and there is no conclusive evidence on the potential tick species associated with the transmission of those Borrelialike organisms.
The authors have not observed dogs with E. canisrelated polyarthritis in their clinical routines and veterinary practices. Dogs from the present study were divided into two groups according to the presence or absence of anti-Ehrlichia spp. antibodies, since the suppression of the immune system in dogs with clinical leishmaniasis could enable the reactivation of a previously subclinical E. canis infection [3], and there can be a synergism between L. infantum and E. canis during co-infection in dogs that can worsen the dog's clinical condition [3]. Although the presence of anti-Ehrlichia antibodies does not confirm an active infection and might merely be indicative of previous exposure to Ehrlichia spp., in the present study it was not possible to perform a molecular diagnosis to confirm Ehrlichia infection status due to financial constraints.
In the present study, the prevalence of CanVL-associated arthritis in dogs was 83.1%, with no difference between groups regarding the percentage of dogs with joint abnormalities on radiography. This prevalence is similar to 91.3% observed in a study conducted with 46 dogs with clinical leishmaniasis independent of the presence of clinical signs of arthritis [13]. Nevertheless, it is much higher than 4.6% [6] and 44.8% [4] observed in two previous studies with dogs with polyarthritis, possibly because radiographic evaluation and SF analysis were performed in eight joints independently of clinical signs of arthritis.
Although some studies have described cases of monoarthritis associated with CanVL [6,9,12,22,23], in the present study all dogs presented involvement of multiple  joints, as has been observed in other studies [4, 7-9, 13, 23]. All evaluated joints were affected bilaterally, but the radiographic changes and the intensity of the inflammatory infiltrate were not the same on both sides. Bone lesions have probably a hematogenous origin, which may explain the presence of bilateral joint involvement [4]. The most affected joints were tarsal (87.8%), followed by carpal (68.9%), elbows (33.7%) and stifles (13.5%), as observed in a previous study in which the same joints were evaluated and the inclusion criterion was the diagnosis of CanVL regardless of clinical signs of arthritis [13]. In other studies in which the inclusion criterion was the presence of clinical signs of arthritis, the joints most frequently affected were the carpal [6] and the carpal and stifles [4].
Radiographic patterns of articular lesions were similar between groups evaluated in present study, with most of the dogs from both groups presenting erosive lesions; these results confirm previous findings [4,6,8,10,11]. Proliferative and mixed (osteolytic and proliferative) lesions were also visualized in both groups, without any evident between-group difference, as also has been reported by other authors [4,7,[9][10][11][22][23][24]. Dogs from G1 were more likely to present soft tissue swelling around the joint (P < 0.0001) and joint space narrowing (P = 0.029); however, we hypothesized that this was just an incidental finding. It has been proposed that subacute or acute stages are associated with non-erosive and chronic stages related to erosive lesions [4].
Although the evaluation of SF in dogs with polyarthritis secondary to Leishmania infection has been previously reported [6,7,9,11,[25][26][27], to the best of our knowledge this is the first study to perform SF evaluation in a large number of dogs (n = 89) with CanVL, with and without previous exposure to Ehrlichia spp., independent of the presence of clinical signs of arthritis. Overall, previous studies have reported an inflammatory infiltrate on SF smears, with either a preponderance of neutrophils [6,9] or mononuclear cells [4,7,11], with Leishmania spp. amastigotes observed on either neutrophils, macrophages or free on the SF [6,7,9,11,[25][26][27]. We found no significant differences in the type of inflammatory infiltrate on SF between joints or between groups. However, there was a slight predominance of mononuclear cells over neutrophils in the inflammatory infiltrates in both groups, as previously reported in a study evaluating SF of dogs with VL [4]. Leishmania spp. amastigotes were found in 27.1% of the joints evaluated, more frequently in joints from G2 dogs (P = 0.002). Although the authors can only assume that G2 dogs had been previously exposed to Ehrlichia spp., it is well known that coinfection between L. infantum and E. canis is common and may alter clinical manifestations [3]. Further studies evaluating SF in dogs co-infected with L. infantum and E. canis should be performed to evaluate this finding.

Conclusions
In conclusion, the present study found a high prevalence of arthritis in dogs with CanVL, with the involvement of multiple joints in all dogs. The most affected joints were tarsal, followed by carpal, elbows, and stifles. There was a slight predominance of mononuclear cells over neutrophils in the inflammatory infiltrates on SF in both groups. Although no difference was observed between groups related to the number of dogs with polyarthritis, most of the radiographic findings and the presence of an inflammatory infiltrate in SF, Leishmania spp. amastigotes were found more frequently in joints from G2 dogs. Further studies evaluating SF in dogs co-infected with L. infantum and E. canis should be performed to evaluate this finding.