Common and atypical presentations of Anaplasma phagocytophilum infection in equids with emphasis on neurologic and muscle disease

1 INTRODUCTION

Granulocytic anaplasmosis in horses (EGA) is caused by an obligate intracellular gram-negative rickettsial organism, Anaplasma phagocytophilum, that infects and replicates in leukocytes.¹ The first report of EGA occurred in northern California in 1969.² The organism previously was known as Ehrlichia equi, hence the name of equine granulocytic ehrlichiosis, but reclassified based on genetic taxonomical analysis.^{3, 4} Cases of EGA have been reported in North America, Brazil, Europe, and North Africa.^{1, 5, 6} The organism is transmitted by ixodid ticks such as Ixodes scapularis and I. pacificus in North America, and I. ricinus in Europe.^7-10 Infection with A. phagocytophilum has been detected in humans and other species worldwide.^11-15 Ticks involved in granulocytic anaplasmosis in humans (HGA) include I. scapularis and I. pacificus in North America, I. ricinus in Europe, and I. persulcatus in Asia.¹⁶

Clinical signs result from vasculitis and vary depending on severity and organ or tissue affected.^{16, 17} After an incubation period of a few days of a tick bite, common clinical signs include fever that can range from 38.9 to 41.7°C (102 to 107 °F), lethargy, anorexia, icterus, mild petechiation, limb edema, stiffness and, in some cases, ataxia.¹⁸ Tachycardia and tachypnea are common during the febrile period, which can last a few days.^{1, 18} Younger animals (<3 years of age) appear to experience milder signs with minimal limb edema.¹ Laminitis and abortion in pregnant mares have not been reported with anaplasmosis.¹

Common laboratory findings include anemia, leukopenia caused by neutropenia and lymphopenia, thrombocytopenia, and hyperbilirubinemia.¹ Inclusion bodies (morulae) in neutrophils or eosinophils are characteristic hematologic findings observed in Giemsa or Wright-stained buffy coat smears from horses and donkeys with EGA.¹ The use of qPCR for the detection of the organism's DNA in buffy coat samples has high sensitivity and specificity.¹⁹ Positive qPCR results for A. phagocytophilum can be detected 2 to 3 days before clinical signs, whereas inclusion bodies in neutrophils can be detected in a buffy coat smear 2 to 3 days after the onset of fever.²⁰

Currently, there are only 2 case studies reporting large numbers of EGA cases (49 and 54 cases) in the literature.^{1, 21} Atypical presentations of EGA occur, as evidenced by sporadic case reports that include horses with rhabdomyolysis, respiratory distress, cavitary effusion, disseminated intravascular coagulation (DIC), recumbency associated with neurologic disease, and death.^22-27 Furthermore, specific detailed descriptions and frequency of neurologic and muscle disease associated with anaplasmosis in equids are scarce in the literature.^{1, 21-24, 28-30} Therefore, our objective was to review retrospectively cases diagnosed with EGA at a referral hospital, to determine the frequency of neurologic and muscle disease, and to describe their clinical signs, laboratory findings, treatment and outcome compared to cases with no neurologic or muscle disease. Other atypical presentations also were reviewed.

2 MATERIALS AND METHODS

3 RESULTS

3.2 Clinical laboratory

Anaplasma phagocytophilum was identified in blood using qPCR (n = 9), cytological analysis (n = 8), or both (n = 22). Clinical laboratory testing consisted of CBC and serum biochemistry panel (Table 1, ST1). Degenerative left shifts with bands were observed in 13 horses with toxic changes in 7 horses, and metamyelocytes were identified in 2 horses. Seven horses with muscle disease had increased of CK and AST activity. Six horses of QH breeds with the N/My allele had median CK activity of 130 229 IU/L (range, 4043-959 453 IU/L; reference range, 119-287 IU/L) and median AST activity of 10 269 IU/L (range, 4222-20 705 IU/L; reference range, 168-494 IU/L). The SDB horse with muscle disease had CK activity of 2763 IU/L and AST activity of 685 IU/L. Kruskal-Wallis one way ANOVA analysis showed significantly higher CK activity in the muscle group (median, 101 077; 95% CI, −112 952, 520 099) when compared to the neuro group (median, 186; 95% CI, 133.1, 364) and no-NM group (median, 160.5; 95% CI, 126.0, 231.9; P = .0002; Figure 1A). Similarly, Kruskal-Wallis one way ANOVA analysis showed significantly higher AST activity in the muscle group (median, 10 269; 95% CI, 3777, 16 081) when compared to the neuro group (median, 258; 95% CI, 231.1, 350.2) and no-NM group (median, 285; 95% CI, 246.7, 424.8; P = .0003; Figure 1B). Other laboratory variables such as packed cell volume (P = .11), white blood cell counts (P = .35), neutrophil count (P = .13), and platelet count (P = .39) were compared among groups and not found to be significantly different. Azotemia was observed in 6 horses (serum creatinine concentration, 2.1-6.1 mg/dL; reference range, ≤2 mg/dL) and considered prerenal and renal in 4 and 2 horses, respectively. Clinical laboratory results are shown in Table 1 (ST1 shows individual animals).

Serum troponin I concentration was measured in 4 horses with suspected cardiomyopathy and found increased (median, 0.27 ng/mL; range, 0.17 to 0.36; reference range, 0 to 0.06 ng/mL). Two of these 4 horses had concurrent neurologic and muscle disease (1 was N/My), and 1 each had neurologic and muscle (N/My) disease. A clotting panel was performed in 3 horses with suspected DIC and found to be abnormal with increased fibrinogen concentration and fibrinogen degradation products (FDP), and decrease of antithrombin III < 50%. Serum amyloid A concentration was measured in 2 horses and found to be increased at 682 and 1774 μg/dL (reference range, 0-7.5 μg/mL), respectively. Urinalysis was performed in 8 horses and found to be abnormal in 2 horses with muscle disease with proteinuria (n = 2, 100 mg/dL) and isosthenuria of 36 hours' duration (n = 1 hydrated horse, USG 1.015).

Abdominocentesis was performed in 4 horses with peritoneal effusion and the fluid was found to be a modified transudate (total protein concentration, 2.5-4.1 g/dL; total nucleated cell count, 5300-13 580/μL). In 1 horse, morulae were observed in neutrophils in the abdominocentesis fluid. Cerebrospinal fluid collected from the lumbosacral space in 2 horses with neurologic disease appeared xantochromic, otherwise cytology was unremarkable (CSF protein concentration and total nucleated cell count within reference range) despite mild numbers of RBC present in the fluid (50 and 190 RBC/μL).

4 DISCUSSION

We described a cohort of 39 equids with anaplasmosis with both common and atypical presentations. Most cases included horses of multiple breeds and 1 donkey. An overrepresentation of males was found in this sample population. Most cases (82%) occurred during the spring and winter months. Clinical signs such as fever, lethargy, anorexia, icterus, and limb edema were observed in equids in our study with 41% having no other manifestations of disease. Neurologic disease was seen commonly (41%) and most often evidenced by symmetrical diffuse proprioceptive ataxia. Other less common presentations included muscle disease (n = 7/39), peritoneal effusion (n = 4), cardiomyopathy (n = 4), and DIC (n = 3) which in some cases occurred concurrently with neurologic or muscle disease. As expected, horses with muscle disease had increases in muscle enzyme activities. Quarter horse breeds with the My allele had severe muscle disease, evidenced by marked increases in muscle enzymes activities. Despite severe signs of disease, Anaplasma phagocytophilum was highly responsive to tetracyclines with signs improving or resolving within hours to days of treatment. Outcome was favorable for all horses, with no recurrence of infection or long-term sequelae.

Similar to humans with HGA, a male overrepresentation was observed in this equine population (humans, 57%-61%; horses, 77%).^{16, 32} Although this finding is of interest, its significance is unclear. In contrast, a sex predisposition was not found in another report of 54 equids.²¹ Experimental infection in pregnant mares results in signs of variable severity with no disruption of fetal development and normal birth of full term foals.¹⁷ Similarly, the 1 pregnant mare in our study recovered fully and the fetus was viable at the time of discharge. A single case report describes natural transplacental infection from a 4-year old Oldenburg mare to a newborn filly that was treated successfully using oxytetracyclines.³³ The median age in our study was similar to that of the other study (14 and 12 years, respectively).²¹ No equids <4 years of age were diagnosed with EGA at our hospital. Young animals (<3 years) have been reported to develop mild disease.^{1, 21} Although naturally-occurring EGA is reported rarely in foals (3 cases in the English literature),^{1, 33, 34} transfer of specific passive immunity to A. phagocytophilum was found in 80% of 22 foals born to seropositive mares with a decrease in antibodies by 3 months of age.³⁵ Despite being rare, the clinician should be prompted to investigate EGA in foals with consistent signs. In people, the median age of HGA is 50 years old.³² Similar to another report from northern California, most cases were seen during the winter and spring months.¹ Also similar to previous studies, low numbers (1 or 2) or no cases were seen during July, August, September, and October.¹ Prevalence of disease might vary depending upon geographical location, because of seasonal weather conditions, environmental temperature, and tick burden. Therefore, cases might present at different seasons of the year in other parts of the United States. Reports of prevalence of disease in a large population of horses in the midwest and eastern United States are lacking.

Clinical anaplasmosis appears to be rare in donkeys based on our study and rare reports of naturally-occurring disease.^{2, 21} However, serum antibodies against A. phagocytophilum have been documented in non-diseased donkeys.^{36, 37} Disease in donkeys has been experimentally induced resulting in signs similar to those described in horses.¹⁷ The sporadic cases of EGA in donkeys and lack of reports in mules suggests resistance or tolerance to natural infection in these species. However, donkeys and mules are underrepresented species at our hospital.

Reported common clinical signs in horses such as fever, lethargy, anorexia, icterus, and limb edema also were observed in our sample population. Previous studies indicated that clinical signs in horses depend on severity and organ or tissue affected with some organs and body regions more susceptible to vasculitis such as testis, ovaries, lung, kidney, brain, heart and limbs.^{2, 17} Diarrhea and laminitis have not been reported in horses with EGA. In our study, neurologic disease was common (41%) and involved the brain (forebrain and brainstem) and spinal cord or both based on neurologic examination by a board-certified veterinary neurologist, with signs being diffuse and symmetrical in most cases. General proprioceptive ataxia was the most common neurologic deficit and ranged from mild to severe with some cases affecting the pelvic limbs more than the thoracic limbs. Severe cases of ataxia can result in recumbency.²⁶ Noting these acute neurologic signs along with fever should prompt the clinician to rule out other acute infectious diseases (eg, equine herpes virus 1 myeloencephalopathy, West Nile, rabies), especially those highly contagious or with zoonotic potential (equine herpes virus 1 myeloencephalopathy, rabies). Multiple cranial nerve abnormalities, tongue paresis, dysphagia, laryngeal dysfunction, and lack of menace were less common in our sample population. These neurologic deficits are rarely reported.²⁴

Abnormalities of clinical laboratory variables vary depending on disease stage and often include anemia, leukopenia with neutropenia and lymphopenia, thrombocytopenia, and hyperbilirubinemia caused by indirect hyperbilirubinemia.^{1, 2} Although these findings also were encountered in our horses, median PCV, WBC, and neutrophils counts were within the reference range in all groups, except for the WBC being below the reference range for the no-NM group (Table 1). Lymphopenia was seen in most horses from all groups. Thrombocytopenia was more commonly seen in horses with no-NM and neurologic disease than in horses with muscle disease. However, the median platelet count was below reference range for all groups. Although hyperfibrinogemia was more severe in horses with neurologic and muscle disease compared to the no-NM group, this difference was not significant. Hyperbilirubinemia caused by indirect hyperbilirubinemia was seen in all groups. Differences in these variables among groups were not significantly different. Horses at our hospital were presented at various stages of disease, which might have accounted for the variability of clinical laboratory test results, especially for those presented with muscle disease. Early diagnosis is essential to initiate appropriate treatment. A previous study reported lymphopenia, hyperbilirubinemia, and hyponatremia as linear predictors, and low creatine kinase activity was a nonlinear predictor of a positive qPCR for A. phagocytophilum.²¹ Hyponatremia and low CK activity were not features in our horses (Table 1) and would have erroneously predicted the results of qPCR. On the other hand, lymphopenia and hyperbilirubinemia were consistent findings in our equine population, except for inconsistent lymphopenia in horses with muscle disease. Therefore, these potential predictors for positive qPCR results do not seem to apply for horses with muscle disease.

Horses with immune-mediated myositis (IMM: N/My, My/My) can have mild leukocytosis (60%) with neutrophilia, and hyperfibrinogemia.³⁰ This observation might explain why horses with EGA in the muscle group had higher WBC, although this difference not significant. Creatine kinase and AST activities were significantly higher in horses with muscle disease compared to other groups. Six of these 7 horses were QH breeds with the My allele. Our results suggest that QH with the My allele might be at risk of developing more severe muscle disease or trigger IMM if infected with Anaplasma. Reports of severe rhabdomyolysis and cardiomyopathy in horses associated with EGA are scarce.²² In our study, 2 horses with IMM also had cardiomyopathy which resolved with corticosteroids and supportive treatment. Two horses with IMM had acute kidney injury, presumed to be multifactorial (dehydration, concurrent administration of non-steroidal anti-inflammatory drugs before referral, myoglobinuria, and possible vasculitis from EGA). These latter horses made a full recovery.

Muscle biopsy results supported a diagnosis of IMM in 2 of 6 QH with the My allele which prompted clinicians to use corticosteroids early in the disease course. Dexamethasone administration at 0.08 mg/kg suppresses the proinflammatory response caused by experimental infection of horses with A. phagocytophilum, resulting in cytokine changes that decrease severity of disease.³⁸ Dexamethasone decreased interferon gamma transcription, interleukin-8, and interleukin-18, and increased interleukin-4.³⁸ Clinically, those horses had delayed development of limb edema, decreased anorexia and icterus, and lower fevers compared to horses not treated with dexamethasone.³⁸ Therefore, the use of corticosteroids in these horses with IMM associated with EGA likely contributed to the rapid favorable outcome. Anaplasma was detected by qPCR in 2 horses with muscle disease. The pathogen infects neutrophils and eosinophils and not muscle, and therefore it is likely that blood contamination and DNA from the pathogen resulted in a positive qPCR. However, no obvious blood contamination or neutrophils were observed histologically in these muscles.

Although uncommon, cavitary effusion involving the pleural and peritoneal spaces associated with EGA can occur and is likely the result of leakage associated with vasculitis, and often is necrotizing.^{17, 25} The observation of morulae within neutrophils from peritoneal fluid has not been reported previously. Disseminated intravascular coagulation is uncommonly reported.²⁷ Based on clinical signs and severity, DIC was investigated and confirmed in 3 horses in our study. Other cases might have been missed because of lack of signs suggesting DIC. Because of the complete resolution of disease, DIC appears to be a secondary transient event.

Similar to humans with HGA, atypical presentations can occur in horses with EGA.¹¹ Here, we described rhabdomyolysis, cavitary effusion, cardiomyopathy, and DIC. Hospitalization often is required in people to manage infection and possible complications such as hemodynamic disturbances, myocarditis, heart failure, renal failure, acute respiratory distress syndrome, central (eg, meningoencephalitis) and peripheral (eg, palsies, demyelinating polyneuropathy, plexonopathy) nervous system disease, acute abdominal syndrome, rhabdomyolysis, and opportunistic infections.¹¹ Although death can occur in people with life-threatening complications, the risk is low (<1%).¹¹ Death has been reported rarely in horses with naturally-occurring and experimentally-induced EGA and associated with comorbidities or complications.^{1, 21, 27} In our study, all equids survived to discharge and recovered fully.

Limitations of our study included its retrospective nature with potential missing or incomplete data. Some horses were treated with non-steroidal anti-inflammatory drugs before referral, affecting rectal temperatures obtained. Horses were admitted at various stages of disease resulting in variability of clinicopathologic data and detection of morulae within neutrophils. Additionally, different tetracycline protocols were used (type of tetracycline, dosage, dosing interval, and duration). However, all equids had favorable outcomes independent of tetracycline protocol with a minimum treatment duration of 6 days. Overrepresentation of atypical presentations is possible because of the nature of the cases seen at a referral hospital. Furthermore, cases presented throughout the year in our study reflect location of northern California; and might be different in other geographical areas.

In conclusion, EGA can manifest with common and atypical presentations affecting males more commonly. Foals and very young animals do not appear or are less likely to be clinically affected. Although rare, donkeys can suffer from naturally occurring disease. Peak months for infection occur during winter and spring in northern California. Neurologic disease was common in horses with EGA presented at our referral institution and mainly manifested as diffuse symmetrical proprioceptive ataxia. Other neurologic manifestations are less common but can occur involving the brain resulting in obtundation and cranial nerve abnormalities such as dysphagia and dyspnea. Although not common, rhabdomyolysis can occur and should prompt clinicians to investigate for MYHM in QH breeds, especially in severe cases. Cardiomyopathy is rare and can occur in horses with and without concurrent muscle disease. Cavitary effusion and DIC are also uncommon but might be missed unless specifically investigated. Depending on stage of disease, leukopenia with neutropenia and lymphopenia are common in horses with EGA, but less common in horses with muscle disease. Thrombocytopenia and hyperbilirubinemia caused by indirect hyperbilirubinemia are consistent findings. Independent of clinical presentation, outcome appears to be favorable regardless of the tetracycline protocol used, and death or long-term sequelae appear to be rare.

CONFLICT OF INTEREST DECLARATION

The authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION

The authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

The authors declare no IACUC or other approval was needed.

HUMAN ETHICS APPROVAL DECLARATION

The authors declare human ethics approval was not needed for this study.