The flagrant neurologic features in this outbreak strongly contrast with those of previous cases, in which acute respiratory disease predominated. However, mild neurologic signs were noted in several horses in 1994 (4), and nonrespiratory manifestations (colic, depression) also have been observed previously (7). Phylogenetic analyses indicate some genetic differences between the strain that caused this outbreak and previously identified strains (K. Halpin, pers. comm.), but experimental infection of horses with the outbreak strain produced both respiratory and neurologic signs (D. Middleton, unpub. data). Virus dose and route of infection could plausibly influence clinical features; alternatively, the neurologic predominance in this outbreak may simply indicate the spectrum of possible manifestations.

Patient records and epidemiologic evidence support our contention that case-patient 1 had the primary case. A roost of Pteropus alecto and P. poliocephalus bats, recognized reservoirs of Hendra virus (9), was located within 5 km of the practice, and bats were regularly observed in the immediate practice vicinity. The reported incubation period for HeV infection in horses is 4–16 days (1,4). We contend that case-patient 1 was exposed to infectious body fluids from a foraging bat (through contamination of pasture, feed, water, or yard rails) and that he was the originator of this outbreak. In the 4 days before his illness and euthanasia, case-patient 1 had been treated at the clinic for a minor laceration. The clinic was the operational hub of the veterinary practice and the evident focus of transmission in this outbreak. We believe that transmission to case-patients 2–4 followed contamination of surfaces or equipment by infectious body fluids (plausibly blood, urine, saliva, or nasal discharge) from case-patient 1 before onset of his clinical signs.

Figure 2

Figure 2. Epidemic curve of Hendra virus infection in horses, Australia, 2008. White bars represent the 3 possible cases; black bars represent the 5 confirmed cases.

Experimental studies suggest that horses may be infectious 48 hours before they show clinical signs (D. Middleton, unpub. data) At the time, case-patients 2–4 were housed in the clinic area or were being intensively treated, with daily or twice daily visits to the clinic, which would likely have increased their probability of becoming infected. Environmental swab samples from stalls 2, 3, and 4 yielded positive results by PCR, confirming that surfaces (walls, doors, wire mesh partitions between stalls) had been contaminated. Case-patient 5 had 3 possible sources of infection: most plausibly, case-patient 4, who had been in the adjacent stall until July 8; less plausibly, the recovered case-patient 2, who had been in stall 2 until July 18; and least plausibly, environmental contamination in stall 3 (occupied by case-patient 3 until July 5). The horse in yard 20, who had an extended opportunity for direct contact with case-patient 1, and the horses in stalls 6 and 7, who had extended opportunity for close contact with case-patients 2 and 5, respectively, did not acquire infection. These findings support previous observations that unassisted horse-to-horse transmission is inefficient. Three other horses died at the clinic without a definitive diagnosis before the death of case-patient 1, but the lack of clinical samples, conflicting epidemiologic evidence, and plausible alternative reasons for death precluded their inclusion as case-patients. However, they are considered possible case-patients (Figure 2).

Results of PCR and serologic tests for all other horses on the premises remained negative 4 weeks after case 5 occurred. No further cases have occurred.

Veterinarians and horse owners should consider HeV infection in any horse exhibiting acute-onset febrile illness, regardless of clinical manifestations, and implement appropriate risk assessment and management strategies. Biosecurity Queensland has compiled comprehensive veterinary guidelines (10)

Two additional spillover events have since occurred: in July–August 2009, 3 horses and 1 human were fatally infected (with a fourth horse nonfatally infected); and in August 2009, two horses were fatally infected.