T. canis, a nematode roundworm of the family Ascaridia, is a cosmopolitan parasite, infecting dogs (and other canids) in all tropical and temperate regions of the world. Toxocariasis in domestic dogs is prevalent almost uniformly in North America south of latitude 60 degrees north and has been reported in all 50 states. The adult worms reside in the proximal small intestine of dogs (the definitive hosts) and live for an average of 4 months. Adult female worms may produce 200,000 eggs/day; eggs passed in feces are not embryonate and thus are not infective. Depending on soil composition, temperature, and humidity, the eggs become infective in 2 to 5 weeks.

In adult dogs, embryonate eggs containing second-stage larvae hatch in the stomach and small intestine, penetrate the intestinal mucosa, travel via the portal circulation to the liver, then enter the systemic circulation, reaching the heart and lungs 3 to 5 days after infection (Box 224.1). Some larvae penetrate the bronchioles, travel to the trachea and pharynx, are swallowed, and develop into adult worms in the small intestine. Other larvae invade the pulmonary vein, travel back to the heart, and spread via the systemic circulation throughout the body. In puppies, the tracheal route predominates, accounting for their importance in the transmission of disease to other hosts.

In humans and paratenic hosts (including mice, rats, lambs, and pigs), the tracheal route of migration leading to the development of adult worms does not occur. Larvae do travel to the liver via the portal circulation and to the systemic circulation via the lungs, however, and they lodge in small blood vessels in somatic organs. The larvae then bore through the walls of the blood vessels and migrate through the tissues. As in dogs, most of these larvae become dormant but may remain viable for many years.

Nearly all human toxocaral infections occur by ingestion of infective eggs from soil that is contaminated with excreta from puppies or from contaminated hands or fomites. Ingestion of uncooked organ and muscle meat from paratenic hosts (pigs, lambs, rabbits, snails, and, perhaps, chickens) is a documented (but uncommon) source of human infection. Pica for dirt (geophagia) is the principal risk factor for VLM in children and adults. Because embryonization requires more than 2 weeks, direct transmission from infected dogs presumably is uncommon. Therefore, frequent exposure to dogs (e.g., by veterinarians) alone is insufficient to predict an increased likelihood of T. canis infection. Although puppy ownership is associated with a higher incidence of T. canis infection, ample exposure may occur in children without a household dog: 10% to 30% of soil samples from public parks, sandboxes, and backyards are contaminated with T. canis eggs, which may survive for years.

Seroprevalence studies using an enzyme-linked immunosorbent assay (ELISA) for antibodies to T. canis have revealed that 4.6% to 7.3% of kindergarten children from different regions of the United States have been infected. Seroprevalence rates are higher in African Americans than in whites. For both African Americans and whites, seroprevalence rates increase with rural residence, crowding, and lower socioeconomic status. In some rural populations in the southeastern United States, seroprevalence rates exceeding 20% have been reported. A positive ELISA for T. canis also is associated with epilepsy, yet children with epilepsy of undefined origin do not have seroprevalence rates higher than those in children with epilepsy of known cause. This finding suggests that epilepsy is a risk factor for the acquisition of T. canis (e.g., through pica) rather than vice versa.

The epidemiologic features of VLM and OLM are strikingly different. Although both are associated with exposure to puppies, only VLM is associated clearly with pica. Patients with VLM usually are 1 to 4 years old, whereas patients with OLM have a mean age of 7 to 8 years. Most patients with OLM have no history of a syndrome similar to VLM, although ocular involvement may occur concomitantly with VLM, especially in very young children with severe disease or many years after VLM.

The clinical and pathologic features of T. canis infection in patients with VLM and OLM reflect primarily the brisk inflammatory response of the host, although the migrating larvae may cause direct tissue damage. Dead or dying larvae provoke a particularly intense inflammatory response. As described in the initial report linking T. canis larvae with VLM, the characteristic pathologic lesions are eosinophilic granulomas that surround larvae in various stages of disintegration; in advanced lesions, no evidence of the larvae remains. Most often in humans, the liver is the site of greatest involvement, but involvement of the lungs also is frequent. Eye involvement is an important complication of T. canis infection and occurs in many different forms. Although they are less common, larval infections of the myocardium, brain, pancreas, skin, kidney, intestine, and regional lymph nodes have been reported.

The contrasting epidemiology of VLM and OLM led Glickman to hypothesize that the dose of the organism ingested could determine whether VLM, ocular involvement, both, or neither developed in the patient. In this model, the ingestion of a few larvae would result in initially asymptomatic infection but could result in ocular disease in some cases. The ingestion of a moderate number of larvae could result in VLM because of a more dramatic inflammatory response; these patients would have a low risk of subsequent ocular involvement if the inflammatory response could prevent migration of larvae to the eye. Finally, ingestion of very many larvae could overwhelm the immune response, resulting in concomitant VLM and ocular disease; these patients would be at higher risk for larval infection of other sites (e.g., brain, myocardium) as well. Although there is some experimental support for features of this model, it remains largely speculative. Three human research subjects given a single dose of 100 to 200 larvae had no clinical evidence of disease but did develop moderate eosinophilia. It is also possible that certain T. canis strains exhibit tropism for ocular or visceral migration.

Experimental infection of paratenic hosts (including mice, rabbits, and the Japanese quail) with embryonate eggs of T. canis has provided important information regarding the pathogenesis of Toxocara infections. These studies have confirmed the importance of the host immune response in the development of tissue injury in this disease. Studies in mice (and in vitro studies of human T lymphocytes) indicate that the T-cell response to Toxocara infection is mediated primarily by cells of the helper-2 T-cell phenotype. Production of the cytokine interleukin-5 (IL-5) by helper-2 T cells appears to be the critical step in the development of eosinophilia during experimental Toxocara infection, and mice genetically deficient in IL-5 fail to develop eosinophilia after challenge with T. canis. Compared with control animals, IL-5–deficient mice exhibit no difference in tissue parasite burden after infection with the embryonate eggs of T. canis, yet they develop less extensive pulmonary damage.