Historical aspects of Potomac horse fever in Ontario (1924–2010)

Abstract

In the summer of 1924 Dr. Frank W. Schofield conducted investigations into an endemic disease of horses in the Kent and Essex counties of Ontario. According to farmers in these counties the disease had existed in this region for at least 50 years previously. The clinical, pathological, histopathological, and epidemiological findings outlined in Schofield’s detailed report strongly suggest that this endemic disease was what was designated in 1979 as “Potomac horse fever” (PHF). This assumption is further substantiated by transmission experiments involving horses and laboratory animals that were conducted by Schofield utilizing horse feces, whole blood, and mayflies. The aim of this paper is to present Schofield’s detailed investigations and findings and to compare these with PHF research conducted from 1979 to 2010 that ultimately led to the discovery of Neorickettsia risticii as the etiological agent and to elucidation of the organism’s complex life cycle.

Introduction

Investigations into an endemic disease of horses in Ontario (1924)

In the 1924 Report of the Ontario Veterinary College, Ontario Department of Agriculture, Dr. Frank Schofield published “An Investigation into an endemic disease of horses (occurring chiefly in Kent and Essex Counties of the Province of Ontario)” (1). On 30 July, 1924, Dr. Schofield began his investigations into this endemic disease that had existed in horses in parts of Kent and Essex counties of Ontario for nearly half a century. Some of the oldest inhabitants at that time stated that as far back as they could remember this disease had existed in that part of the country. This disease was known locally by farmers as “cholera,” “horse cholera,” “dysentery,” and “abdominal typhoid” (1).

Season of the year

Schofield reported that “scarcely a year passes without the occurrence of the disease, sometimes in a very acute form with high mortality; in other seasons the infection is not severe but the number of cases unusually large”. The farmers in the affected areas considered “the disease as very serious due to its constant occurrence in the haying season.” The dangerous period in Kent and Essex counties was from the middle of July to the beginning of August. In 1924 the first case occurred on July 14 (1).

Distribution of the disease

The disease appeared to be confined to the counties of Kent and Essex, which are the most southern parts of Ontario. This part of the province is a peninsula, bounded on the north by Lake St. Clair, on the west by the Detroit River, and on the south and east by Lake Erie. In his report, Schofield stated “there is a definite relationship between the water front and the occurrence of the disease”. He also stated that “in general, one might say that the zone of infection extends from the water front to a distance of about five miles inland”. The disease was almost unknown in the town of Essex, which is one of the farthest inland points of the peninsula. The most seriously affected areas were between the towns of Amherstburg and Harrow in the south, and around Mitchell’s Bay in the north. Another interesting comment that Schofield made was “that the disease rarely occurs in horses that are living in the towns of the infected zone.” “During the 1914–1924 period only one or two cases occurred in the town of Amherstburg while in the surrounding territory there have been several hundred cases” (Figure 1) (1).

Figure 1.

Map of Counties of Essex and Kent, Ontario. The dotted areas indicate the chief centers of the disease affecting horses in the summer of 1924 (modified from reference 1).

Prevalence and mortality

Although it was not possible for Schofield to obtain reliable information as to the number of cases of this disease prior to 1924, in 1916 “a severe epidemic” was reported. During July, 1924, an extensive and serious outbreak of the disease occurred in Kent and Essex counties involving approximately 130 cases with 22 deaths (1).

Description of the disease

The severity and sometimes even the signs of the disease were “modified from year to year.” The most constant characteristic of the disease was an acute diarrhea accompanied by a marked depression. In the acute form the initial signs noted were “dullness and marked prostration”. The animal was lethargic and anorexic. When made to walk the gait was uncertain or even staggering. The animal appeared to be very “toxic” (injected, dirty brownish-red ocular mucous membranes). All visible mucosae were injected. The limbs were constantly cold, while the back and buttocks were either hot or cold. The temperature was elevated to 103°F to 105°F. The pulse rate was elevated (70 to 90 beats/min) and was “full and firm.” “Palpitation of the heart” was not uncommon. The respiratory rate was always rapid and markedly abdominal. Intestinal borborygmi were increased with “loud rumbling and gurgling sounds” which could “sometimes be heard at quite a distance”. Quite early in the course of the disease diarrhea almost always occurred, and this was regarded as a favorable prognostic sign. When diarrhea was absent or only slight the farmers called the disease “dry cholera” and this was associated with a less favorable prognosis than “wet cholera.” Sometimes very frequent bowel movements occurred and the feces were described as “like turbid brown water”. In the mild form of the disease the signs observed in the acute form were present but to a lesser degree. Frequently the disease was not considered until the horse exhibited diarrhea. The diarrhea lasted for a few days and was not accompanied by the severe depression and systemic signs observed in the acute form. The horse’s appetite soon improved and there was only a slight loss of body weight. An interesting feature of the disease was that the animals usually never showed acute colic and rarely showed mild signs of colic. Laminitis was a common sequel in both the acute and mild forms of the disease (1).

Duration and progression of the disease

The “crisis” with this disease typically occurred on day 4 to day 5 of the illness. In the mild form body temperature dropped to normal, appetite returned, and the animal quickly recovered. In the acute form the horse continued to be pyrexic and the diarrhea became more pronounced or no feces were passed (“obstinate constipation”). Affected horses would not eat, were very depressed, and rapidly lost weight. There was little or no evidence of colic, but affected horses frequently shifted their feet. The respiration became more rapid and shallow, the pulse fast and weak, the lungs became congested, and the animal became recumbent a short time before dying with little or no struggle. Death usually occurred from 7 to 10 days after the onset of the disease (1).

Clinical pathology

No abnormalities were found in 6 cases on examination of blood films stained with Wright’s stain. A white blood cell count was performed on only 1 horse. In the acute stage this was 8.0 × 10⁹/L and 8.1 × 10⁹/L during convalescence. Numerous blood and fecal cultures were performed on affected horses but no specific organism was isolated (1).

Necropsy findings

The postmortem findings on 3 horses that died in July 1924 in Kent and Essex counties were reported by Dr. Schofield. One horse died on day 8 and another died on day 15 after the first signs of illness. The duration of the disease was not stated in the third case.

Case 1

There was marked congestion of the subcutaneous capillaries and the blood was thick and dark. There was no fluid in the abdominal cavity and the viscera had a markedly dry appearance. A fibrinous exudate was present on the visceral surface of the bowel. The intestine was friable and easily torn. The major findings were an acute and extensive inflammation of the mucosa of the large and small intestine. The inflammation of the small intestine involved the mucous membranes of the duodenum and many areas of the jejunum and ileum. The mucosa of the cecum, portions of the large colon and “floating colon” was acutely inflamed and was a dark red to almost purple. A small quantity of a chocolate-colored slime was present in areas where the inflammation was most pronounced. The mesenteric lymph nodes were dark red and edematous. The liver was light in color and friable. The spleen appeared somewhat shrunken and had numerous petechiae. The kidneys were congested. In the thoracic cavity the lungs were congested and filled with dark, tarry blood. The myocardium was light in color and had the appearance of boiled beef.

Case 2

The subcutaneous vessels were engorged with quite a number of hemorrhages in the sub-peritoneal fat. Almost the entire serosal surface covering the cecum was rough and dull due to a fibrinous exudate. Throughout the length of the small intestine there were numerous large and small areas of catarrhal inflammation. The lungs were congested and the heart had several ecchymoses under the endocardium and pericardium. The myocardium was light in color and only a few areas of normal heart muscle remained.

Case 3

Acute and extensive inflammation of the mucous membranes of both small and large intestine was present.

The gross postmortem findings in these cases strongly indicated to Schofield the presence of an acute general infection (1).

Histopathological findings

The main histopathological findings of 3 fatal cases occurred in the small and large intestine. The more superficially placed epithelial cells were entirely removed and even the deeper cells of the crypts frequently showed signs of degeneration and separation from the basement membrane. The intestine was infiltrated by large and small lymphocytes and plasma cells. The kidney tissue was badly damaged and showed cloudy swelling and necrosis of the epithelium of the convoluted tubules (1).

Epidemiology

Schofield (1) observed that cases did not develop from one common source of infection and there was not much evidence that the disease spread from 1 sick horse to another. He noted that the first 4 cases in the 1924 outbreak occurred on 4 different farms on 4 consecutive days (July 14 to July 17). Although the first 2 cases occurred on adjoining farms there was no known contact and he considered that 24 h was an insufficient incubation period. He noted that the disease was not easily transmitted from animal to animal and that it was unusual for more than 1 case to occur in the same stable. In the Harrow district 15 cases occurred on 13 farms. Only 2 farms reported 2 cases (1).

Etiology

After considering all the available evidence, Schofield was convinced the disease was due to microbial infection. This was despite the fact that no constant bacterial species was isolated on culture of blood, feces, or tissue. His reasons for considering that this disease was due to infection were: (i) the clinical picture was suggestive of infection (pyrexia, tachycardia, tachypnea, with marked depression); (ii) the pathological changes seen in tissues were commonly seen in acute infections; and (iii) the fact that 1 attack of the disease almost always conferred lifelong immunity. The evidence suggested to Schofield that there were either multiple sources of infection, possibly each farm possessing its own focus of infection, or a common source of infection which was distributed by insects or some other carrier of disease. Schofield also stated that one must not overlook the possibility of microbial growth in the pasture. It was not possible to associate the disease with any kind of soil, or with changes in the rainfall and temperature (1).

The role of the mayfly (Ephemerida) in this disease

Schofield (1) stated that “this ephemeral insect was found in unusual numbers in the district where the disease is most prevalent. The flies breed in the lake and river and when the wind is off the water the mayflies are carried over the land like swarms of locusts. They are seen piled up inches deep under the electric lights, and may be so thick in the pasture that it is impossible to walk without treading on numbers of them. Now, the strange thing is that the disease always appears a few days after the May flies appear, and is at its worst by the time the May flies disappear”. In further discussion Schofield states: “Unquestionably the most popular belief as to the cause of “horse cholera” is the “May fly theory”. No one knows how the May fly causes so serious a malady, but most are quite sure that it does. The chief objection to the “May fly theory” is that the May fly occurs in many parts of Ontario where the disease is unknown. Also cases of this disease have occurred long after the May fly has disappeared. In favor of the “May fly theory,” one must point out that the flies are more prevalent along the water front than inland. The flies, like the disease, decrease as the distance from the water increases. Some experiments have already been made with May flies, but more must be undertaken before such a popular theory is discarded. It may be mentioned in passing, that thus far the only factor common to all cases is the May fly. However, the number of flies ingested by stabled horses must have been small” (1).

Transmission experiments

Experiments were conducted by Schofield to determine whether the disease was transmissible to horses and to animals of other species (rabbits, guinea pigs). His hypothesis was that if transmissibility could be proven then a living agent could be postulated as the cause of the disease. The first horse experiment (Experiment a) was to determine whether the disease was transmissible from horse to horse. Two horses were orally dosed with approximately a quart (946 mL) of feces made by mixing the diarrheal feces from 3 well-marked cases of the disease. As well, approximately 60 mL of citrated blood obtained from 3 horses with acute disease was administered intravenously to 2 horses. On day 6 of the experiment 1 of the 2 horses showed a typical but mild form of the disease. The horse was anorexic, pyrexic (40°C), tachycardic (60 beats/min), tachypneic (60 breaths/min), and had diarrhea for 2 days. The animal recovered but remained weak for some time. The other horse showed no signs of illness.

The second horse experiment (Experiment b) was to determine whether the disease could be produced by the ingestion of mayflies. Two horses were orally dosed with a suspension of ground mayflies. The material consisted chiefly of moltings rather than the bodies of dead mature flies. The molts of several hundred flies with many dead flies were given to each horse. No untoward effect was observed. Schofield commented that it was unfortunate that live flies could not be obtained for the experiment (1).

Schofield (1925) then conducted experiments in laboratory animals. In Experiment 1, 4 rabbits were injected intravenously with 4 mL of citrated blood from horses with acute disease. Two rabbits died within 7 d of the administration of blood. Postmortem examination of both cases showed impaction in the large intestine, with localized peritonitis. Feces from the 2 dead rabbits was fed to healthy rabbits — both rabbits remained perfectly well. In Experiment 2, 3 guinea pigs were fed large quantities of dead mayflies. The mixture chiefly consisted of the molts of maturing flies and the carcasses of dead flies that had been collected from the vicinity of electric lights. One of the 3 guinea pigs died 5 d later and the postmortem examination showed catarrhal inflammation of the small intestine. A healthy guinea pig that was fed the feces from this dead guinea pig also died 7 d later and showed identical lesions, with marked inflammation of the gall bladder as well. No specific organism was isolated on bacteriological culture of the feces and tissues of these 2 guinea pigs. Large numbers of an amoeboid organism were found in the intestinal mucus but no such organism was found in the feces of diseased horses.

In Experiment 3, 8 guinea pigs and 3 rabbits were fed samples of feces from typical cases of “horse cholera.” Two of the 8 guinea pigs died, 1 on day 4 and the other on day 5 after the ingestion of the feces. The postmortem findings closely resembled those seen in the guinea pigs that were fed the may-flies. One of the 3 rabbits died showing intestinal inflammation.

The transmissibility experiments of Schofield came to a sudden end due to the marked decline in new cases and the termination of the endemic (1).

Prophylaxis

The only prophylactic measure that Schofield advised was the stabling of horses and feeding on old hay during the dangerous period from the middle of July to the beginning of August. It appeared that stabled horses, which were hand fed, were not as susceptible as horses running at pasture. Schofield concluded that the disease should be designated “acute intestinal catarrh” as it differed from ordinary forms of intestinal catarrh in that it was endemic, had a high mortality, and produced profound depression that was not accompanied by severe pain in the affected horses (1).

Potomac horse fever (PHF) (1979–1997)

In 1979, “Potomac horse fever” (PHF) was first recognized as a distinct clinical entity by veterinarians in Montgomery County, Maryland, USA (2–5). The disease occurred in horses on farms in an area adjacent to the Potomac River in Montgomery County, Maryland, and Fairfax and Loudon counties, Virginia (2,3,6–8). Between 1979 and 1981, most PHF cases in Maryland were seen “within a mile or two of the Potomac River” in a geographical strip approximately 6 miles long (2). Between 1982 and 1986, the number of clinically recognized and reported cases in this region reached 904 (9). The dramatic increase in incidence and focal nature of the disease led to speculation that this was a new disease. Initially the disease was called “Acute Equine Diarrhea Syndrome (AEDS)” (2). The term “Potomac horse fever” was coined by a television reporter covering the original epizootic along the Potomac River (8). At first, it was believed that PHF was limited to a small region along the Potomac River in Maryland and Virginia; however, the disease was subsequently recognized by veterinarians in other states (2,7,10).

Season

Epidemiologic studies have shown that PHF is markedly seasonal in occurrence with a concentration of cases in July, August, and September (2–4,6–9). Cases may occur from May to November (2,4,11); however, almost 70% of the 904 cases recorded in Maryland between 1982 and 1986 occurred during July and August (7,9).

Clinical signs

Clinical signs of PHF vary markedly and may include any or all of the following: fever, depression, anorexia, ileus, mild to severe diarrhea, colic, edema of the distal limbs, and laminitis with approximately 30% mortality (2,3,7,8,10–13). Any combination of these signs may be present, with only the rare case showing most of them (8).

Initially, horses with PHF show an acute onset of depression, and anorexia, followed by fever (38.9°C to 41.7°C). Decreased borborygmi in all abdominal quadrants is often noted initially and is followed in 24 to 48 h by higher pitched tinkling sounds suggestive of excessive fluid and gas accumulation (2–4,8). Diarrhea is common but does not occur in all cases (2,3,13). Fecal consistency may be normal on the initial examination but within 24 to 48 h a moderate to severe diarrhea develops in approximately 60% of affected horses. In some horses the diarrhea is transient; in others it may persist for up to 10 d with fecal consistency ranging from “cow-pie” to profuse, watery in character (2,3). The onset of diarrhea is often accompanied by mild colic signs (3). Some horses develop severe toxemia and dehydration, which results in cardiovascular compromise that is characterized by elevated heart and respiratory rates and congested mucous membranes. Subcutaneous edema may be observed along the ventral abdomen and limbs. Laminitis occurs in 15% to 30% of PHF cases and usually occurs within 3 d of the initial diarrhea. Laminitis may progress despite the resolution of other clinical signs (2–4). The course of the disease without therapeutic intervention is usually 5 to 10 d (8). The mortality rate of clinical cases of PHF ranges from 17% to 36%. Some PHF horses have to be euthanized as a result of secondary complications, such as acute or chronic laminitis (2–4,6–8).

Clinical pathology

Leukocyte counts are highly variable, ranging from leukopenia to normal or leukocytosis (11,14). In experimentally induced PHF the most significant hematologic change is a leukopenia (< 5.0 × 10⁹/L) that develops between days 8 to 19 post-infection in 80% to 90% of horses. This is followed by a leukocytosis (> 14.0 × 10⁹/L) that develops in approximately 50% of the horses between days 13 to 30 post-infection (13). No organisms are seen on a standard blood smear (11). Clinical biochemistry results are non-specific (11).

Epidemiology

“Epizootics” may occur in endemic areas; however, the disease is non-contagious with a sporadic pattern of occurrence (3,4,7,9,10). Within an endemic area, many animals develop PHF; however, large outbreaks do not necessarily occur on any 1 farm (4). Following the initial reports of PHF in Maryland and Virginia (2,7), other enzootic areas were identified in the United States (10). In all but one of these areas the focus of disease was along a major river (4,10). In 1991 a retrospective serological study revealed that the disease had been present in the United States “for many years”. Geographic variations in seropositivity to Neorickettsia risticii were observed in New York state. Higher risk counties were regions of low elevations that contained or bordered large bodies of water or large river valleys (15). Serological data indicate that in North America, PHF now occurs in 43 states of the United States, and 3 provinces of Canada (Ontario, Alberta, Nova Scotia) (16).

Transmission

Experimental reproduction of PHF was reported in 1984 following the intravenous (IV) injection of whole blood from infected horses or ponies to susceptible ponies (3,17). The minimal amount of heparinized blood necessary to transmit PHF is not known but ranges from 20 to 400 mL (3,14,18–20). The incubation period between the time of inoculation and development of fever and leukopenia varied from 8 to 15 d (2,3,18). Over a 2-week period only 1 of 8 ponies orally challenged with 6 doses of approximately 1 kg of fresh feces from infected ponies developed signs consistent with PHF. The other 7 remained asymptomatic, and were susceptible to IV infection (5).

Etiology

In 1984, researchers used transmission electron microscopy to demonstrate a rickettsial organism in the large colon of a pony experimentally infected with PHF (18). The round shape and location of the organism in vacuoles of macrophage-like cells suggested that it belonged to the genus Ehrlichia in the Family Rickettsiaceae (18). Further studies revealed the occasional presence of ehrlichiae in the small colon, jejunum, and cecum (21). The microorganism was present in the cytoplasm of deep glandular epithelial cells, mast cells, and in macrophages migrating between glandular epithelial cells in the lamina propria and submucosa (22). The causative organism was isolated from the leukocyte fraction (23) and from cultured blood monocytes (24) of experimentally infected ponies. The disease was then experimentally reproduced by inoculating a susceptible pony with culture material (19,24,25).

The causative organism was named Ehrlichia risticii(12,24). In 2001 the bacterium was placed in the genus Neorickettsia and renamed Neorickettsia risticii(26). Neorickettsia risticii is a gram-negative coccus that stains dark blue to purple with Giemsa and Romanowsky’s stains. The agent can be grown in cell culture but cannot be isolated by conventional bacterial culture (16). The causative agent was isolated in pure culture from whole blood (27). Oral transmission of PHF was demonstrated in 2 of 4 ponies following the nasogastric intubation of Ehrlichia-infected mouse monocyte tissue culture cells sealed in gelatin capsules (28). In a later transmission experiment, 6 of 9 ponies inoculated with N. risticii via nasogastric intubation and oral drench developed PHF and seroconverted (5). Potomac horse fever has also been called “equine ehrlichial colitis (EEC)” (8,10,23,29,30), “equine monocytic ehrlichosis (EME)” (13,15,19,12,29–34), “Shasta River Crud” (16,35), “churrio” or “churrido equino” (36). The disease has been documented in France (37) and the Netherlands (38).

Diagnostics

Diagnostic tests for the detection of antibodies or DNA from N. risticii have been developed (19,39). An indirect fluorescent antibody (IFA) test was developed in 1986 (19). Serology is not a straightforward diagnostic technique because of the kinetics of antibody production stimulated by N. risticii(8). Infected horses have a rapid rise in antibody titer, which usually begins before the onset of clinical signs. Antibody levels peak quickly reaching a high point in a few days (8). A 4-fold change in titer over a period of 3 to 4 wk is considered diagnostic (4). Detection of a significant rise in IFA titer is considered impossible in the absence of an early serum sample. The IFA test titer does not distinguish between present and past infection or vaccination and single serum samples may be of limited value because horses may carry a titer of 1:640 for a year or longer (4,19,40). A one-step polymerase chain reaction (PCR) can detect DNA of N. risticii in the blood (39) and feces of experimentally infected horses (41). In naturally infected horses, detection of N. risticii DNA by a nested PCR was more effective with peripheral blood than with fecal specimens (40). From clinical cases of PHF, N. risticii was detected by PCR in the blood of 81% (17 of 20) of culture-positive blood samples, and is considered as sensitive as culture (40).

Potomac horse fever in Ontario (1995–2010)

The first reports of PHF in Ontario were published in 1995 involving horses in southwestern (42) and eastern (29) Ontario. During the summers of 1993 and 1994, PHF was diagnosed in 4 horses in eastern Ontario. Diagnosis was made on the clinical signs and seroconversion based on a 4-fold increase in IFA titer between acute and convalescent sera. All 4 horses were native to Ontario and had been kept in different stables in Carleton, Stormont, Grenville, and Leeds counties (29). In July, 1995, PHF was diagnosed in 4 horses on 3 farms on the shores of Lake St. Clair in Chatham-Kent County (42). Acute and convalescent serum samples were collected from 3 cases and subjected to the IFA test. Although no seroconversion was demonstrated, the convalescent titers ranged from 1:2560 to 1:10240 (42). Between 1995 and 2010 PHF has been confirmed in 20 horses at the Ontario Veterinary College Veterinary Teaching Hospital. Confirmation was based on clinical signs, season, location, IFA seroconversion, PCR on blood and/or feces, and histopathological findings if the horse died. The distribution of these cases is shown in Figure 2.

Figure 2.

Location of laboratory confirmed cases (n = 18) of Potomac horse fever presented to the Ontario Veterinary College Veterinary Teaching Hospital (1995 to 2010). The dots represent the origins of cases.

Life cycle of Neorickettsia risticii

Neorickettsia risticii is an obligate intracellular bacterium of trematodes and mammals and the reservoir for the organism is most commonly a trematode species with 2 life stages in intermediate hosts (30,43,44). In its first stage, the trematodes parasitize freshwater stream snails (44–46). In the natural environment, Neorickettsia spp. can be trans-stadially transmitted through all developmental stages of trematodes and trans-ovarially passed through generations of trematodes (43). The relationship of N. risticii with its trematode host seems to be commensal or mutualistic, as reproduction of trematodes does not appear to be adversely affected by infection (33,47,48). Mammalian infection by Neorickettsia spp. occurs by horizontal transmission of the bacterium from trematodes to susceptible mammalian hosts, mostly through ingestion of this bacterium in the metacercarial stage of trematodes encysting in insects or fish (33). In the eastern United States, N. risticii is maintained by the digenetic trematode, Acanthatrium oregonense, which has a complex life cycle consisting of miracidia and sporocysts in snail hosts (Elimia virginica), free-swimming cercariae, metacercariae in aquatic insects (caddis flies, mayflies), and adults that lay eggs in the intestinal lumen of insectivorous bats (33,47, 49,50). Aquatic insects living in close association with snails in stream or lake water are infected with N. risticii-carrying trematodes and transmit both the trematodes and the rickettsial agent to potential hosts such as fish, amphibians, birds, reptiles, and mammals (43). During periods of warm water temperatures, cercaria infected with N. risticii are released from the snails, infecting and developing into metacercaria in the second intermediate host (aquatic insects). Aquatic insects living in close association with snails were found to be infected with N. risticii-carrying trematodes and were considered likely to play an important role in the epidemiology of PHF (43). Using nested PCR amplification and sequence analyses, N. risticii has been detected in metacercariae found in immature and adult caddisflies (Trichoptera), mayflies (Ephemeroptera), damsel flies (Odonata, Zygoptera), dragonflies (Odonata, Anisoptera), and stone flies (Plecoptera). Overall the prevalence of N. risticii in aquatic insects was 31.9% (n = 454 individuals) with 15.2% prevalence in mayflies (n = 92) (51,52).

Horses develop PHF when they ingest aquatic insects containing encysted N. risticii-infected trematodes. Upon ingestion of N. risticii in the metacercarial stage of the trematodes in aquatic insects by horses, N. risticii is horizontally transmitted from the trematodes to horses and replicates within inclusion bodies inside monocytes, macrophages, mast cells, and intestinal epithelial cells (21,44,53,54). Infected intestinal epithelial cells lose microvilli, which may contribute to the reduced electrolyte transport and water resorption. An increase in intracellular cyclic AMP is found in infected horse intestinal tissues and this may also contribute to reduced luminal absorption of Na⁺ and Cl⁻ in the colon, and thus to the lack of water absorption and diarrhea (33,55). Neorickettsia risticii cells are shed into the intestinal lumen (22) and are found in the feces (33,44). Koch’s postulates concerning the natural transmission of PHF in horses were fulfilled by feeding infectious snail secretions and mature aquatic insects (i.e., caddisflies) (44,51,52). Clinical signs developed 10 to 15 d after ingestion (44,51,52). Mayflies were implicated in a cluster of PHF cases in horses in Minnesota and Iowa within 3 wk of their attendance at a horse show in southeastern Minnesota. During this horse show, competitors reported that vast numbers of mayflies blew into the facility and horse trailers at the show grounds, contaminating the stalls, hay supplies, and water of many horses. Positive identification of N. risticii was made through PCR in dead mayflies inside the show facility (56).

Mayfly biology relevant to Lake St. Clair and Lake Erie

Burrowing mayflies, Hexagenia spp., can readily be found in shallow bays, basins, and connecting channels of the Laurentian Great Lakes. Two regions where they have been historically abundant are Lake St. Clair and western Lake Erie, which are characterized by conditions favored by Hexagenia; that is, both are relatively shallow, have extensive soft bottom regions, and are mesotrophic (57). In the Laurentian Great Lakes, Hexagenia species normally spend 1 to 2 y in the nymphal stage (58). Final-instar nymphs leave the substrate, swim to the water surface, and moult into a subimago. Subimagos emerge from the water and fly or are carried by wind to land where they rest for a day and then moult into a sexually mature imago (59). Aquatic insects such as mayflies typically have a relatively short period to disperse, and exposure to meteorological conditions likely influences flight significantly (59). Mayflies are relatively poor flyers but Hexagenia adults fly or are carried inland by wind an average of 1.2 km (59).

Conclusions

This review describes the investigation of an endemic disease of horses in Kent and Essex counties in Ontario conducted by Dr. Frank W. Schofield in the summer of 1924. Based on Schofield’s detailed clinical findings, geographical location of cases, proximity to lakes and rivers, and time of the year in which horses were affected, the authors conclude that the disease that he was investigating was what is now designated “Potomac horse fever” (“equine monocytic ehrlichiosis,” “equine ehrlichial colitis”). This conclusion is further substantiated by transmission studies Schofield conducted utilizing feces of affected horses and mayflies. CVJ