On October 2, 2009, a mare in Kleberg County, Texas, USA, showed clinical signs of equine piroplasmosis. Serologic testing at the Animal Plant Health Inspection Service, National Veterinary Service Laboratories (NVSL), US Department of Agriculture (USDA) (Ames, IA, USA) with a commercially available competititve ELISA (VMRD Inc., Pullman, WA, USA) detected T. equi antibodies. The remaining 359 horses on the index ranch were tested in the same way, and 292 (81.1%) of 360 were seropositive for T. equi on initial screening (Table 1).

Ticks collected from horses on the index ranch were shipped alive to NVSL. Identifications were made by using morphologic characteristics, geographic distribution, biologic characteristics, and host associations (5–9). NVSL received ticks from 228 horses; >1 species was present on 41 animals. The dominant tick, Amblyomma cajennense, was collected from 180 (78.9%) horses (Table 2).

All ticks were identified and sent to the Agricultural Research Service, Animal Disease Research Unit, USDA (Pullman, WA, USA) for transmission studies. Live males and partially fed females were pooled by species and held at 25°C and a relative humidity of 98% until they were allowed to reattach and feed on uninfected horses.

A total of 104 A. cajennense ticks (45 male and 79 female) were placed on a horse on October 30, 31, and November 2, 2009. These ticks had been removed from 73 horses on the index ranch, of which 68 (93.2%) were seropositive for T. equi. Females were allowed to reattach and feed until repletion; males were removed when all females were replete. All ticks were removed by November 18, 2009. Twenty-four fully engorged females and 3 live males were recovered. The horse had a fever (>39°C) 14 days after the ticks were first applied. Parasitized erythrocytes on a stained blood smear peaked at 0.3% on day 17. No other clinical signs of infection were evident. Serologic analysis and PCR (10) confirmed T. equi infection.

Twenty-nine D. variabilis ticks (12 male and 17 female) were placed on a second uninfected horse on October 30 and 31 and November 2 and 12, 2009. These ticks had been removed from 17 horses, of which 11 were seropositive and 1 was seronegative for T. equi; 5 had an unknown infection status. All ticks were removed by November 24, 2009. Six fully engorged females and 7 live males were recovered. This horse had a slight fever (39°C) 15 days after tick attachment but otherwise showed no clinical signs. No organisms were found in blood smears, but this horse was positive for T. equi by PCR 42 days after the first ticks were attached and by competitive ELISA 87 days after tick attachment.

Ranch staff reported that they used no practices that would result in movement of blood-contaminated materials between horses (e.g., no reuse of needles), which suggests that iatrogenic transmission was not responsible for this outbreak. Consequently, the high prevalence of T. equi infection implies a focus of vector-borne transmission.

A. cajennense ticks were the most abundant species on horses during the period (October–November) of this investigation (Table 2). Our results demonstrate that A. cajennense ticks naturally acquired infection while feeding on infected horses and transmitted T. equi intrastadially when they reattach and feed on uninfected hosts. A. cajennense ticks have not been shown experimentally to be a competent vector for T. equi (1). This species is a 3-host tick, and all life stages are known to feed aggressively on a wide variety of hosts, including horses. The natural distribution of A. cajennense ticks includes southeast Texas; they are not known to be present in other parts of the United States where cases of equine piroplasmosis have occurred (11). Although this study demonstrates that A. cajennense ticks are an experimental intrastadial vector, additional studies are needed to fully characterize the vector capacity of this species, particularly with regard to interstadial transmission.

Immature stages of D. variabilis ticks occur almost exclusively on rodents; only adults were found on horses at the ranch. Although these ticks were able to transmit T. equi intrastadially to an uninfected horse, the small proportion of infested horses on the ranch (16.2%; Table 2) and low transmission efficiency of this species (G.A. Scoles, unpub. data) make it unlikely that D. variabilis ticks were responsible for the high infection prevalence.

A. maculatum, the Gulf Coast tick, was the second most abundant species on horses at the index ranch during the study (19.7%; Table 2). However, this species survived poorly during handling and transport, probably because it is less tolerant of desiccation than are A. cajennense ticks (12), and we did not have enough viable ticks to attempt transmission feeding. Whether this species can act as a vector for T. equi is unknown. D. (Anocentor) nitens ticks were collected from 7 (3%) horses sampled. This species is a proven vector of Babesia caballi but has not been shown to be a vector of T. equi. R. microplus ticks are limited to a quarantine zone along the Texas–Mexico border. The index ranch is north of this zone, and no R. microplus ticks were found on horses at this ranch.

Although T. equi can be transmitted iatrogenically, e.g., by common needle use (13), this route of transmission is improbable with good management practices. Vector-borne transmission is more likely than iatrogenic transmission to establish and maintain a large focus of infection, such as in this outbreak. Additional tick studies are needed to determine whether other indigenous tick species are involved in transmission at this site. However, if A. cajennense ticks are the primary vector, the outbreak will likely be confined to this region because southeastern Texas is the northern extent of the range of this tick in the United States (11). Given knowledge of tick species that are competent vectors, spread of this parasite can be controlled by testing requirements and limits on regional movement of equines on the basis of presence or absence of such competent vectors.