Potomac horse fever (PHF) is an acute enterocolitis syndrome producing mild colic, fever, and diarrhea in horses of all ages, as well as abortion in pregnant mares. The causative agent is Neorickettsia risticii. The infection of enterocytes of the small and large intestine results in acute colitis, which is one of the principal clinical signs of PHF. The disease is seen in spring, summer, and early fall and is associated with pastures bordering creeks or rivers. The epidemiology of PHF has been shown to involve a trematode vector. Sporadic disease caused by N risticii has been reported in dogs and cats; cattle appear to be resistant to infection. PHF has been reported in many areas of the USA and Canada using an indirect fluorescent antibody test as evidence of exposure; however, recent studies indicate a high rate of false-positive titers with this test, and the true geographic range of distribution is not known. Isolation or detection of the causative agent from clinical cases of PHF using conventional cell culture or PCR assay has been reported only from California, Illinois, Indiana, Kentucky, Maryland, Michigan, New York, New Jersey, Ohio, Oregon, Pennsylvania, Texas, and Virginia.
N risticii is a gram-negative obligate intracellular bacterium with a trophism for monocytes. Initial morphologic studies of this organism isolated from cell culture, as well as the serologic responses of N risticii, caused this bacterium to be assigned to the genus Ehrlichia. However, DNA analyses have shown N risticii is most closely related to N helminthoeca, the agent of salmon poisoning in dogs, and Ehrlichia sennetsu, a disease of people in Japan. The organism is not visible in monocytes in blood films from clinical cases, in contrast to Anaplasma phagocytophilum, which is readily identifiable in granulocytes of infected horses.
N risticii has been identified in freshwater snails and isolated from trematodes released from the snails. N risticii DNA was detected in 13 species of immature and adult caddisflies (Trichoptera), mayflies (Ephemeroptera), damselflies (Odonata, Zygoptera), dragonflies (Odonata, Anisoptera), and stoneflies (Plecoptera). Transmission studies using N risticii–infected caddisflies have reproduced the clinical disease. One route of exposure is believed to be inadvertent ingestion of hatched aquatic insects that carry N risticii in the metacercarial stage of a trematode. The incubation period is ~10–18 days. The causative organism is present in the feces of experimentally infected horses, but the biologic significance of this is unknown. Clinically ill horses are not contagious and can be housed with susceptible horses. Additional studies are needed to determine the exact role of the vector and helminth hosts in the complex maintenance cycle of N risticii.
The clinical features of PHF are typified initially by mild depression and anorexia, followed by a fever of 102°–107°F (38.9°–41.7°C). At this stage, intestinal sounds may be decreased. Within 24–48 hr, a moderate to severe diarrhea, with feces ranging in consistency from that usually seen in cows to watery, develops in ~60% of affected horses. The onset of diarrhea is often accompanied by mild abdominal discomfort. Some horses develop severe signs of sepsis and dehydration. Clinical signs can be indistinguishable from those of Salmonella and other infectious causes of enterocolitis. Laminitis can supervene as a severe complication of PHF in 20%–30% of affected horses. Hematologic findings vary in the early stage of PHF from leukopenia (characterized by neutropenia and lymphopenia) and thrombocytopenia to a normal hemogram, despite evidence of systemic illness. A common finding in cases of PHF is a marked leukocytosis, which is normally seen within a few days of onset. PHF may present with all or any combination of these clinical signs.
Several months after clinical disease in pregnant mares, abortion due to fetal infection with N risticii may occur. Experimentally, pregnant mares infected at 100–160 days of gestation abort at 190–250 days of gestation. The abortion is accompanied by placentitis and retained placenta. Fetal lesions include colitis, periportal hepatitis, and lymphoid hyperplasia of mesenteric lymph nodes and spleen. Necropsy findings in nonpregnant horses with enterocolitis are nonspecific and reveal diffuse inflammation, mainly in the large intestines.
A provisional diagnosis of PHF often is based on the presence of typical clinical signs and on the seasonal and geographic occurrence of the disease. A definitive diagnosis should be based on isolation or identification of N risticii from the blood or feces of infected horses by cell culture or PCR. Serologic testing is of limited value as a diagnostic tool, although many infected horses have high antibody titers at the time of infection. Because of the high prevalence of false-positive titers, interpretation of the indirect fluorescent antibody test in individual horses is difficult. Rising paired titers can be helpful. Isolation of the agent in cell culture, although possible, is time-consuming and not routinely available in many diagnostic laboratories. A real-time PCR assay that allows detection of N risticii DNA within 2 hr is a much more feasible test for routine diagnostic examination. To enhance the chances of detection of N risticii, the assay should be performed both on blood and fecal samples, because the presence of the organism in blood and feces may not necessarily coincide.
Horses with PHF can be treated successfully with oxytetracycline (6.6 mg/kg, IV, bid), if given early in the clinical course of the disease. A response to treatment is usually seen within 12 hr. This is associated with a drop in rectal temperature, followed by an improvement in demeanor, appetite, and borborygmal sounds. If therapy is begun early, clinical signs frequently resolve by the third day of treatment. Generally, antimicrobial therapy is for no more than 5 days. In animals that exhibit signs of enterocolitis, fluids and NSAIDs should be administered. Laminitis is more common than in other causes of enterocolitis, and if it develops, is usually severe and often refractory to treatment. The overall case fatality rate is 5%–30%.
Several inactivated, whole-cell vaccines based on the same strain of N risticii are commercially available. Although vaccination has been reported to protect 78% of experimentally infected ponies, it has been marginally protective in the field. Vaccine failure has been attributed to antigenic and genomic heterogeneity among the >14 different strains of N risticii isolated from naturally occurring cases. Furthermore, vaccine failure may also be due to lack of antibody protection at the site of exposure, because the natural route of transmission has been determined to be oral ingestion of the agent. Minimizing insect ingestion in stabled horses by turning off barn lights at night, which normally attract the insects, has been suggested.
No zoonotic risk is known.