|Home | About | Journals | Submit | Contact Us | Français|
The raccoon roundworm, Baylisascaris procyonis, is the most common and widespread cause of clinical larva migrans in animals. In addition, it is increasingly recognized as a cause of devastating or fatal neural larva migrans in infants and young children and ocular larva migrans in adults. Humans become infected by accidentally ingesting infective B. procyonis eggs from raccoon latrines or articles contaminated with their feces. Two features distinguish B. procyonis from other helminthes that cause larva migrans: (i) its aggressive somatic migration and invasion of the central nervous system and (ii) the continued growth of larvae to a large size within the central nervous system. Typically, B. procyonis neural larva migrans presents as acute fulminant eosinophilic meningoencephalitis. Once invasion of the central nervous system has occurred, the prognosis is grave with or without treatment. To date, despite anthelmintic treatment of cases of B. procyonis neural larva migrans, there are no documented neurologically intact survivors. Epidemiologic study of human cases of neural larva migrans demonstrate that contact with raccoon feces or an environment contaminated by infective eggs and geophagia or pica are the most important risk factors for infection. In many regions of the United States, increasingly large populations of raccoons, with high rates of B. procyonis infection, live in close proximity to humans. Although documented cases of human baylisascariasis remain relatively uncommon, widespread contamination of the domestic environment by infected raccoons suggests that the risk of exposure and human infection is probably substantial. In the absence of early diagnosis or effective treatment, prevention of infection is the most important public health measure.
Baylisascariasis in humans is caused by infection with the nematode parasite Baylisascaris procyonis. Baylisascaris procyonis and related species are large nematodes of the order Ascaridida. Other, more familiar ascarids are Ascaris lumbricoides, Toxocara canis, and Toxocara cati, nematode parasites of humans, dogs, and cats, respectively. Baylisascaris species are primarily parasites of lower carnivores. The North American raccoon (Procyon lotor) is the definitive host for B. procyonis. Baylisascaris procyonis is considered the most common cause of clinical larva migrans in animals, in which it is usually associated with fatal or severe neurological disease. It is distinguished from other causes of larval migrans by its propensity for aggressive somatic migration, larval invasion of the central nervous system (CNS), and the capability for continued larval growth within intermediate hosts. More recently, the zoonotic potential of B. procyonis has become evident. In its most severe form, B. procyonis is a rare cause of fatal or neurologically devastating neural larva migrans (NLM) in infants and young children. Characteristically, B. procyonis NLM presents as acute eosinophilic meningoencephalitis. Epidemiologic studies suggest that pica or geophagia and exposure to infected raccoons or environments contaminated with their feces are the most important risk factors for infection. To date, despite treatment, neurological outcome is dismal in the overwhelming majority of documented cases. However, most cases of B. procyonis infection are preventable by relatively simple measures. This review examines the epidemiologic, pathogenetic, clinical, and diagnostic features of baylisascariasis in humans and emphasizes the importance of increasing public awareness and ongoing preventive efforts.
Baylisascaris procyonis was first isolated (as Ascaris columnaris) from raccoons in the New York Zoological Park in 1931 (38). It was later recognized as a distinct species (Ascaris procyonis) in raccoons in Europe (61). The genus Baylisascaris was defined by Sprent in 1968 as including eight recognized and two provisional species previously classified as members of the Ascaris or Toxascaris genus (60). The new genus was named after H. A. Baylis, formerly of the British Museum of Natural History, London, United Kingdom.
The propensity of B. procyonis larvae to produce fatal NLM in various experimentally infected and wild rodents was recognized by Tiner from an early stage (62). More recently, B. procyonis was recognized as the most common and widespread cause of clinical larva migrans in animals, being capable of producing severe or fatal NLM in over 100 species of birds and mammals (28, 29, 44).
The possibility of human infection was anticipated by Beaver (4) and later by Kazacos and colleagues (32). The marked zoonotic potential of B. procyonis has become apparent only in the last 2 decades. The first confirmed cases of NLM in humans were described to have occurred in two young boys, in 1984 and 1985 (16, 25). Since that time, another 11 confirmed human cases have been documented (Table (Table1)1) (8, 10, 18, 41, 44, 47, 51). An earlier (1975) unconfirmed case of NLM involved an 18-month-old girl with geophagia from Missouri, who presented with acute hemiplegia, cerebrospinal fluid (CSF) and peripheral eosinophilia, and elevated isohemagglutinins (1). Serological testing, which was not available for Baylisascaris at that time, was positive for Ascaris and negative for Toxocara in serum and CSF. Because Ascaris infection is not characterized by somatic or CNS migration, those authors astutely suggested that another animal ascarid species more closely related to A. lumbricoides than to T. canis was the most likely cause. Baylisascaris procyonis fits this description well. The closely related species B. columnaris and Baylisascaris melis, parasites of skunks and badgers, respectively, are rare causes of clinical larva migrans in animals and, potentially, in humans (29).
Baylisascaris procyonis infection in raccoons is usually subclinical; adult nematodes are confined to the small intestine (Fig. (Fig.1)1) (29). Adult B. procyonis organisms are large, tan-colored roundworms; the female is larger (20 to 22 cm long) than the male (9 to 11 cm) (Fig. (Fig.2).2). In the raccoon intestine, adult female worms produce prodigious numbers of eggs, with estimates of between 115,000 and 179,000 eggs/worm/day (29). In nature, infected raccoons shed an average of 20,000 to 26,000 eggs per gram of feces and can shed in excess of 250,000 eggs per gram of feces (29). Thus, infected raccoons can shed millions of B. procyonis eggs daily, leading to widespread and heavy environmental contamination.
Individual raccoons may have a heavy parasite burden (mean burden, 43 to 52 worms). Juveniles have a significantly higher prevalence of infection (93.5%) than adults (55.3%), the highest rate of egg shedding, and consequently a greater potential for environmental contamination (58). The higher parasite burden of juvenile raccoons (mean burden, 48 to 62 worms) than in adults (mean burden, 12 to 22 worms) likely reflects differences in mechanisms of infection (13, 43, 58). Young raccoons are infected in the first few months of life by ingestion of infective eggs (containing second-stage larvae), which stick to their mother's fur or contaminate the den and its surroundings (29, 43). Adult raccoons, in contrast, ingest third-stage larvae during predation or scavenging of infected intermediate hosts.
Baylisascaris procyonis eggs are ellipsoidal and dark brown and measure from 63 to 88 μm by 50 to 70 μm in size (Fig. (Fig.3)3) (29). The eggs contain a large single-celled embryo surrounded by a thick shell with a finely granular surface. Eggs of the closely related Toxocara species are lighter in color, have a coarsely pitted shell, and are slightly larger in size (85 by 75 μm). Baylisascaris procyonis eggs are not immediately infective after shedding. With suitable environmental temperature and humidity, B. procyonis eggs usually develop into infective second-stage larvae in 2 to 4 weeks and, occasionally, as early as 11 to 14 days (29, 52). In nature, B. procyonis eggs resist decontamination and environmental degradation and remain viable in moist soil for years (29).
After ingestion by young raccoons, second-stage larvae hatch from infective eggs in the small intestine. Larvae then migrate to the intestinal mucosa, where they develop before they reemerge into the lumen and undergo final development into adult worms (mean interval, 63 days) (29). After ingestion by adult raccoons, third-stage larvae encysted in tissues of intermediate host prey also hatch in the small intestine, where they remain within the lumen and mature into adult worms (mean interval, 35days) (29).
In nature, intermediate hosts, usually small birds and mammals (particularly rodents) are infected by ingesting infective B. procyonis eggs while foraging for food at preferred sites of raccoon defecation, termed latrines. After ingestion by intermediate hosts, infective eggs hatch in the small intestine, larvae rapidly penetrate the intestinal mucosa, migrate via the portal circulation through the liver to the lungs, gain access to the left side of the heart via the pulmonary veins, and are distributed to the tissues by the systemic circulation (29, 33, 59, 63).
A small but potentially devastating number of larvae (typically 5 to 7%) then enter the CNS (33). Aggressive migration and growth of larvae, particularly within the CNS, leads to the debilitation or death of the intermediate host. Rodents, rabbits, primates, and birds appear most susceptible to NLM following ingestion of infective B. procyonis eggs (29). Larvae eventually become encapsulated within eosinophilic granulomas, where they remain viable until they are ingested by raccoons or for the lifetime of the host (33). Raccoons are omnivorous and opportunistic carnivores, preying on debilitated intermediate hosts and consuming larvae encysted in their tissues. Humans are accidental intermediate hosts. Infection typically occurs in young children with pica or geophagia after ingestion of infective B. procyonis eggs from environments or articles contaminated with raccoon feces (30).
A number of reports have documented infection of domestic dogs and puppies with egg-laying adult B. procyonis worms (2, 22, 29). Because of their close contact with humans, particularly children, B. procyonis infection of domestic dogs and pets represents a greater potential risk of infection and is a worrisome development. In addition, because dogs defecate indiscriminately, the potential exists for even more extensive contamination of the domestic environment. At present, descriptions of infection of domestic pets by adult B. procyonis are limited to experimental studies and isolated case reports (40). The prevalence of infection of domestic dogs is unknown but may be more widespread than is appreciated. Unless eggs or worms shed in dog feces are examined closely, B. procyonis may be misidentified as the closely related Toxocara or Toxascaris species.
To fully understand baylisascariasis in humans, it is necessary to consider the key epidemiologic role played by the raccoon. Raccoons are common free-ranging mammals, native to the Americas from Canada to Panama (27, 37). Baylisascaris procyonis is indigenous in North American raccoons and has been demonstrated to occur in raccoons introduced around the world. The prevalence of B. procyonis infection is high in wild raccoons in Germany and those kept in zoos or as pets in Japan (3, 29, 40, 54). Raccoons were formerly introduced into Europe (France, Germany, and The Netherlands), the Soviet Union, and Asia for the commercial fur trade and into Japan as pets (27, 29).
The risk of animal and human exposure to B. procyonis is related to numbers and the prevalence of infected raccoons and latrine density in a given locale. In North America, raccoons are extremely common in rural, suburban, and urban settings, where they have become well adapted to living alongside people. Raccoons thrive in areas with a permanent water supply, a readily available source of food, and suitable sites for dens (24, 27). Some of the highest raccoon population densities are described to be in and around suburban and urban parks and residences (24, 47, 48). The abundance of supplemental food (from gardens, garbage, bird feeders, and pet food) and den sites and reduced predation and hunting in these areas further contribute to increasingly high raccoon population densities (48). Stomach contents of wild raccoons are frequently found to contain (43%) pet food, suggesting that such food is made accessible to raccoons, either intentionally or inadvertently (42). The predominantly nocturnal habits of raccoons may conceal the presence of these large populations (Fig. (Fig.4).4). Thus, in many parts of the United States, large populations of raccoons with a high prevalence of B. procyonis infection now live in close proximity to humans (18, 42, 47, 50). According to the Department of Natural Resources, in 1999, raccoons in the United States were responsible for in excess of an estimated $41 million in damage (15). In the absence of hunting and trapping, it is projected that raccoon populations in the northeastern United States will double over the next decade (34).
In North America, the prevalence of B. procyonis is high in the Midwest, northeast, and West Coast raccoon populations, where prevalences of 68% to 82% are reported (highest in Illinois) (14, 29, 47). However, over time, the local prevalence of B. procyonis in raccoons varies and is subject to change because of natural migration, translocation to restock hunting areas, or accidental transport by garbage trucks (29, 67). Thus, veterinarians, physicians, and public health officials need to be alert to the possibility of zoonotic Baylisascaris infection outside of traditional high-risk areas. Specifically, B. procyonis is an emerging infection in raccoons in the southeastern United States, an area traditionally considered as being low risk (12). It seems likely that B. procyonis infection is a possibility wherever raccoons are found.
Although generally solitary animals, raccoons defecate in preferred communal sites, termed latrines (19, 27). Raccoon latrines play a central role in the transmission dynamics of B. procyonis (44). Individual latrines are visited by many raccoons and may contain substantial quantities of feces; a single latrine may contain between 370 and 750 g of feces (46, 50). Active raccoon latrines are characterized by the constant addition of fresh feces to older decomposing feces, such that eggs found at raccoon latrines are at various stages of development (46). Latrines are routinely visited by a variety of small mammals and birds, which forage for undigested seeds and grain (46). In nature, most transmission of B. procyonis to intermediate hosts occurs by ingestion of infective eggs from raccoon latrines (46). Similarly, contact with raccoons or their feces is the most important risk factor for infection in cases of B. procyonis NLM in humans (18, 29).
In forested and rural areas, raccoon latrines are characteristically found on raised flat surfaces, such as stumps and limbs of trees, large logs, downed timber and rocks, and at the bases or in the crotches of trees (Fig. (Fig.5)5) (18, 24, 46, 50). In suburban and urban settings, raccoon latrines are also found in lofts, attics and chimneys, and on flat roofs, wooden decks, woodpiles, and patios (Fig. (Fig.5)5) (24, 29, 43, 46, 50). In addition, soil near latrines is frequently heavily contaminated with infective eggs and constitutes a potential long-standing source of infection for foraging animals and children with geophagia or pica. Over 3,300 infective B. procyonis eggs were recovered from a single 20-g soil sample collected adjacent to an active raccoon latrine (18). Typically, eggs in soil surrounding raccoon latrines originate from older degraded feces and commonly contain infective second-stage larvae. Human cases have also resulted from egg-laden raccoon feces in roof latrines being washed down gutters by rainfall and contaminating the surrounding ground below (44, 50, 53).
Raccoon feces are typically dark, tubular, and 7 to 15 cm long by 2 cm in diameter, have a pungent odor, and may contain a variety of foodstuffs, such as corn fragments, undigested seeds, bones of small mammals and fish, shells of crustaceans, and remnants of garbage (24, 44). Old feces may resemble leaves or debris. Just as raccoon latrines are an attractive and important source of food for many animal and bird species, they conceivably may represent attractive play areas for young children (30, 46). Because B. procyonis eggs may remain infective for years, long after surrounding raccoon feces has degraded, contaminated areas can serve as long-term sources of infection for susceptible animals and humans (29).
The first confirmed cases of human B. procyonis NLM were described in 1984 and 1985 in 10- and 18-month-old boys from Pennsylvania and central Illinois, respectively (16, 25). Both presented with rapidly progressive and ultimately fatal eosinophilic meningoencephalitis. Other reports followed, namely, of severe nonfatal NLM in young boys from upstate New York (n = 1), California (n = 3), Illinois (n = 2), and Michigan (n = 1) and in a young adult with developmental delay from Oregon (n = 1) and of additional fatal disease in toddlers from Minnesota (n = 2) and a teenager with developmental delay from California (n = 1) (Table (Table1)1) (8, 10, 18, 41, 44, 47, 51). Notably, raccoon populations in these areas have the highest prevalence of B. procyonis infection (29). Within these regions, cases have occurred in rural, urban, and suburban areas.
In epidemiologic studies, two factors placed these patients at risk for severe infection: (i) contact with infected raccoons, their feces, or a contaminated environment and (ii) geophagia or pica. Specific risk factors included exposure to baby or pet raccoons; indoor storage of contaminated downed timber, wood chips, or bark for firewood; indoor contamination by raccoon dens in chimneys and fireplaces; or outdoor contamination of children's play areas by raccoon feces. In field studies, infective B. procyonis eggs have been isolated from bark and wood chips, soil from play areas, and surrounding latrines. All patients were observed eating dirt or debris or sucking wood chips from contaminated areas or were known to exhibit pica or geophagia. Pica or geophagia is common in infants under 2 years of age (7). In addition, young infants and older developmentally delayed patients have relatively poor hygiene. Presumably, if exposed to a contaminated environment, such individuals are more likely to ingest large quantities of B. procyonis eggs and are predisposed to severe NLM.
Of the documented cases of NLM, all were males, and 11 of the 13 cases occurred in toddlers or young children. The male predominance is probably related more to play habits rather than an inherent increased susceptibility to infection. Conceivably, raccoon latrines may make particularly attractive play areas for inquisitive infants and children (30). Several of the cases had an antecedent diagnosis of developmental delay or mental handicap. The only documented cases in older patients occurred in 17- and 21-year-old males with developmental delay and geophagia living in areas of Los Angeles and Oregon, respectively, with high raccoon populations (8, 10).
Baylisascaris procyonis is also increasingly recognized as a cause of ocular disease in humans. In contrast to NLM, which is almost exclusively restricted to infants and young children, isolated ocular larva migrans (OLM) usually occurs in otherwise healthy adults (21, 31). Most commonly, in such cases there is no obvious or only incidental exposure to raccoons or their feces and no history of pica. This suggests that isolated OLM follows the ingestion of relatively few B. procyonis eggs and the chance migration of a single B. procyonis larva into the eye.
The apparent increase in cases of NLM and OLM in humans parallels the increase of cases in animals and of raccoon populations and the contamination of domestic environments in the same areas. The contributions of increased awareness of the condition and of the availability of a diagnostic test to this observed “emergence” are unclear.
The full clinical spectrum of human baylisascariasis is unknown but includes visceral larva migrans (VLM), NLM, and OLM. In addition, preliminary evidence suggests that asymptomatic infection also occurs. As in animals, it is likely that the clinical presentation of B. procyonis infection in humans is determined primarily by the number of eggs ingested, which in turn determines the number of larvae entering the CNS. Additional determining factors likely include the site and extent of larval migration in the CNS and size of the host brain (29). Thus, severe NLM is almost exclusively described to occur in young infants with geophagia or pica who are most likely to ingest large numbers of eggs. In addition, their relatively small brain size increases the likelihood that larval migration will cause severe clinical disease. In contrast, in adults with OLM, contact with raccoons or their feces is incidental or not apparent, the number of eggs ingested is presumably smaller, and the brain size is larger. The absence of eosinophilia or positive antibaylisascaris serology in OLM supports this hypothesis (21, 31).
The finding of antibaylisascaris antibodies in healthy asymptomatic adults, presumably reflecting exposure to small inocula of eggs and the presence of isolated VLM or a small number of larvae in silent areas of the larger adult brain, suggests that low-level infection may be more common than is appreciated (9, 10).
Larva migrans is defined as the prolonged migration and persistence of helminth larvae in the organs and tissues of humans or animals (4, 29). In severe infections of infants and young children, as in intermediate animal hosts, B. procyonis larvae undergo widespread somatic migration. Numerous granulomata have been demonstrated histologically in the heart, mediastinal soft tissues, pleura and lungs, small and large bowel walls, and mesentery and mesenteric lymph nodes (16, 25). Nonspecific clinical manifestations of this migration include macular rash (predominantly on the face and trunk), pneumonitis, and hepatomegaly (10, 16, 18, 25, 51). It is presumed that the development of dyspnea and tachypnea, which is seen in primates 3 to 5 days after experimental infection, is secondary to early pulmonary migration (32, 33). Baylisascaris procyonis is also considered the likely but unproved cause of the sudden unexpected death of a mildly mentally handicapped 10-year-old boy from Massachusetts who presented with an inflammatory polypoid mass in the left ventricular myocardium and marked peripheral eosinophilia (5). At autopsy, the intracardiac mass consisted of several large (60 to 70 μm in diameter) degenerating ascarid larvae surrounded by eosinophilic granules and granulomata. Additional visceral or CNS involvement was not apparent, and confirmatory serologic testing was not performed.
The incubation period of baylisascariasis in humans is unknown; however, if the inoculum of eggs is large enough, fulminant NLM can develop within 2 to 4 weeks. Although a relatively low percentage of ingested larvae enter the CNS (5% to 7% in mice), widespread migration and continued growth within host tissues leads to extensive damage (29).
The majority of human patients with B. procyonis NLM present with an acute fulminant eosinophilic meningoencephalitis. Common early features include low-grade fever, ataxia, increasing lethargy, somnolence, and periods of increased irritability. Over time there is regression and loss of developmental milestones, progression to extensor posturing, increasing spasticity with hemi- or quadriparesis, and ocular or cranial nerve involvement. Seizures occur commonly and may be difficult to control. Neurologic status may deteriorate rapidly to stupor, coma, and death. To date, all survivors have been left in a persistent vegetative state or with severe residual deficits. Blindness and visual impairment are common sequelae, and most patients require total nursing care.
A more subacute, indolent encephalopathy has occasionally been described to occur in older patients with developmental delay and geophagia. This presumably reflects the accumulation over time of a sufficient number of larvae to produce clinical disease within a larger-size brain (10, 18).
Ocular disease in baylisascariasis occurs in association with severe NLM or as an isolated finding. Most infants and children with clinical VLM and NLM also have evidence of ocular disease. Here, visual impairment or blindness results from widespread larval migration, with destruction of the visual cortex, or from larval migration within the eye itself. Ophthalmoscopic exam demonstrates choroidioretinitis, optic neuritis, or atrophy and occasionally may reveal motile larvae migrating within the retina (21, 31, 39). Morphometric measurement of retinal larvae allows differentiation of the larger B. procyonis larvae (1,500 to 2,000 by 60 to 70 μm) from those of related Toxocara spp. (350 to 445 by 20 μm), which are a somewhat more common cause of isolated OLM (21).
Baylisascaris is now considered the most common cause of the large nematode variant of diffuse unilateral subacute neuroretinitis (DUSN), a form of OLM associated with progressive monocular visual loss and changes in retinal pigmentation and optic nerve anatomy (21, 39). Severe localized ocular damage results from inflammation of the retina, retinal vasculature, and optic nerve in response to the local presence of B. procyonis larvae (21, 31).
Although well characterized for animals, the pathology of human baylisascariasis is based on autopsies of the first two fatal human cases (16, 25). In these cases, the brain was the most severely affected organ. Massive larval invasion of the CNS is characteristic, with estimates of more than 3,200 larvae being isolated from the brain of a single case (3 larvae/g of tissue) (16). Acute, macroscopic findings included marked swelling and softening of the brain, leptomeningeal congestion and thickening, and evidence of cerebellar herniation (Fig. (Fig.6)6) (16). Necrosis was most marked in the inner third of the periventricular white matter, with numerous “track-like” spaces being visible to the naked eye. In contrast, generalized atrophy, with thickening of the basal and spinal meninges, was evident in a child who died 14 months after acute infection (25). Severe generalized atrophy was evident in the cerebral hemispheres, corpus callosum, basal ganglia, thalami, cerebellum, brain stem, and spinal cord, in association with a loss of gray-white matter demarcation and ventricular enlargement.
Histologic findings for acute fatal cases demonstrate necrosis and inflammation with numerous macrophages, eosinophils, lymphocytes, and occasional plasma cells concentrated in cerebral periventricular white matter and leptomeninges. Characteristically, large numbers of eosinophils and eosinophilic granules and deposits of extracellular eosinophilic material (the Splendore-Hoeppli phenomenon) are present around necrotic migration tracks and cerebral blood vessels (23). Immunofluorescence studies suggest that eosinophilic material surrounding migration tracks, larvae, and perivascular areas consists of eosinophil major basic protein arising from eosinophil degranulation (23). Moertel et al. and others demonstrated markedly increased CSF levels of eosinophil major basic protein, eosinophil-derived neurotoxin, and interleukin 5, a critical cytokine involved in eosinophilia and the eosinophilic response to helminth parasites, in two fatal cases of B. procyonis NLM (35, 41, 49). At autopsy, larvae have been identified in cerebrum, cerebellum, and spinal cord, with and without surrounding inflammatory reaction. In one of the first fatal cases, 185 live motile larvae were recovered postmortem from one 60-g sample of cerebrum (16).
Subacute cases of NLM demonstrate well-defined granulomata, composed of relatively few intact eosinophils surrounded by a chronic fibrotic reaction, and the absence of major basic protein deposition, consistent with subsidence of the acute inflammatory reaction (25). Microscopically, CNS granulomata containing larvae surrounded by a chronic fibrotic reaction were concentrated in the deep periventricular white matter.
In fatal human cases, NLM is accompanied by widespread visceral involvement. Somatic larval migration causes mechanical tissue damage and necrosis and provokes an intense eosinophilic and granulomatous inflammatory reaction. Macroscopically, VLM manifests as multiple small whitish nodules 1 to 1.5 mm in diameter on pleural surfaces, soft tissues surrounding the laryngo-tracheo-bronchial tree, lung hila, epicardium and myocardium, the wall of the colon and ileocecal regions, small bowel mesentery, and mesenteric lymph nodes (16, 25). Microscopically, the nodules consist of granulomata containing coiled B. procyonis larvae surrounded by macrophages, eosinophils, plasma cells, and dense fibrous collagenous tissue (16, 25).
Histologically, B. procyonis larvae are 60 to 80 μm in greatest diameter and have a characteristic appearance, best seen in transverse section through the midbody/midintestinal region (29). Characteristic diagnostic features include a large, centrally located, and laterally compressed intestine; paired triangular lateral excretory columns flanking the intestine, and prominent single lateral cuticular alae on either side of the body (Fig. (Fig.77).
Although OLM is described in conjunction with NLM, descriptions of ocular pathology are lacking (25, 39, 47, 51). In naturally and experimentally infected animals, B. procyonis OLM is characterized histopathologically by the presence of larval migration tracks in the retina, progressing to necrosis and intense eosinophilic inflammation of the retina, choroid, and vitreous (29, 31). The clinical description of intra- and subretinal motile larvae of the size of B. procyonis in cases of NLM or OLM suggests that acute pathology in humans is similar (21, 31). Histologic findings in the late stages of DUSN demonstrate retinitis, retinal and optic nerve perivasculitis and atrophy, peripheral retinal degeneration and depigmentation, and low-grade choroiditis (17).
Humans are accidental intermediate hosts for B. procyonis. Infection follows the ingestion of B. procyonis eggs containing infective second-stage larva. In humans, as in natural intermediate hosts, B. procyonis larvae undergo aggressive tissue migration that includes the CNS and eyes. Although clinical NLM is a common sequela of infection, B. procyonis larvae are not inherently neurotropic. Experimental and wild animal data and human autopsies suggest that CNS involvement results from extensive somatic larval migration. Continued larval growth and migration following entry of even a few B. procyonis larvae into the CNS may have potentially severe consequences. In marked contrast, the related nematode Toxocara, a much more common cause of VLM and OLM, is a very unusual cause of NLM (55, 64).
Animal and limited human autopsy data suggest that B. procyonis larvae hatch in the small intestine, penetrate the bowel mucosa, and pass, presumably via the portal circulation, through the liver and along vascular channels to the lungs (16, 25, 29). Although hepatomegaly is described for baylisascariasis, hepatitis and hepatic larvae or granulomata are not prominent. In contrast, Toxocara larvae commonly affect the liver, as larvae are trapped once the host becomes sensitized (55). In the lungs, B. procyonis larvae rupture pulmonary capillaries, enter pulmonary veins, return to the left side of the heart, and gain access to the systemic circulation. From the systemic circulation, larvae may be distributed throughout the body, including the CNS. Larvae most likely access the brain by penetrating cerebral blood vessels. After hatching in the intestine, a small number of larvae migrate locally in the intestinal wall, mesentery, and lymphatics and move from the lungs to the pleura, hilar tissues, mediastinum, and heart.
Parasite and host factors contribute to the success of B. procyonis as a parasite and to the pathogenesis of baylisascariasis in humans. Parasite factors include the fecundity of adult worms, the longevity and durability of the ova, the broad host range, the ability of the larvae to continue to grow to a large size within the intermediate host, and the toxicity of larval excretory and secretory products. Host factors include the size of the host brain and the host eosinophilic inflammatory response.
The long persistence and resistance of B. procyonis eggs to environmental degradation or efforts at decontamination are the principal reasons for the success of B. procyonis as a parasite. The eggs are extremely hardy, and given adequate moisture they can remain viable and potentially infective in soil for many years, long after latrines have ceased to be active and surrounding raccoon feces has degraded (29). Larvae even remain viable in eggs stored for months in 10% formalin (44). In addition, B. procyonis eggs tend to be sticky and adhere to animal fur, surfaces, and objects, including children's toys and presumably the fingers of infants. The combination of prolific egg laying by adult B. procyonis worms and the persistent viability of eggs further increase the potential for larval transmission.
Like most other ascarids, B. procyonis has no obligatory intermediate host (43). However, the broad host range of B. procyonis larvae is most unusual and likely represents a distinct survival advantage for the parasite (29). Specifically, the ability of larvae to infect over 100 species of animals and birds increases the number of potential intermediate hosts for the parasite. Aggressive somatic migration that includes the CNS and continued growth of larvae within the CNS also confer a survival advantage and are likely to have been selected for over time (29). Neural larva migrans that causes death or debilitation of intermediate hosts increases the probability that the life cycle is perpetuated as larvae are transmitted back to predatory or scavenging raccoons (29, 62). Baylisascaris procyonis larvae demonstrate somatic tropism for the cranial rather than caudal regions of intermediate hosts. Thus, B. procyonis VLM is characterized by preferential concentration of larvae in the head, neck, and thoracic regions (32, 56, 63). Distribution of larvae to these areas probably occurs by migration from the left side of the heart via the first major arterial branches of the systemic circulation arising from the aortic arch (32). This preferential somatic distribution may create a survival advantage for B. procyonis; raccoons appear to first eat the head, neck, and muscular regions of their prey and frequently discard the rest of the carcass (56, 63). Again, concentration of larvae in these areas probably increases the chances of their being ingested by raccoons.
The relatively large size of B. procyonis larvae also contributes to the pathogenicity of this organism. Baylisascaris procyonis larvae molt and continue to grow and develop as they migrate, reaching comparatively large size (Fig. (Fig.7)7) (29). Specifically, second-stage larvae measure approximately 300 μm in length when they hatch from ingested eggs. During their somatic migration in intermediate hosts, they grow and develop into third-stage larvae measuring 1,500 to 1,900 μm in length by 60 to 80 μm in maximal diameter (16, 21, 52). Migration of such large larvae within the CNS is likely to exacerbate mechanical tissue damage. Among helminths that cause VLM, such continued growth of larvae is unusual; e.g., larvae of Toxocara spp. measure 350 to 445 μm in length and do not grow within intermediate or paratenic hosts (45, 55). And finally, migrating B. procyonis larvae also leave a trail of highly antigenic and neurotoxic secretory and excretory products in their wake (29). These larval products cause marked local tissue necrosis and recruit a characteristically strong eosinophilic host inflammatory response (23, 41). Eosinophil degranulation releases eosinophil proteins (including major basic protein, eosinophil-cationic protein, and eosinophil-derived neurotoxin), which are toxic to mammalian cells, in vitro and in vivo (20, 35).
Unfortunately, the host inflammatory response to B. procyonis larvae is neither protective nor curative and is probably damaging. Despite a vigorous serum and tissue eosinophilic response, encapsulation of larvae within eosinophilic granulomata is slow and cessation of migration is delayed, particularly within the CNS. The finding of larvae in normal-appearing brain without surrounding inflammation and with only a few intact eosinophils suggests that the inflammatory reaction lags behind actively migrating larvae (23). In addition, Baylisascaris larvae remain viable while encapsulated in tissues of human and intermediate animal hosts (16, 59). When the inflammatory reaction eventually develops, eosinophil degranulation causes extensive necrosis of surrounding tissues and is itself neurotoxic (23). Furthermore, survival of B. procyonis larvae in eosinophilic granulomata within animal and human tissues suggests a role for additional mechanisms to evade the host immune response (16, 59).
Finally, epidemiologic and experimental animal data support the hypotheses that infants and young children are at especially high risk for development of severe NLM because of their propensity to ingest large numbers of infective B. procyonis eggs and because of their relatively small brain size.
The combination of encephalopathy with CSF and peripheral eosinophilia and diffuse white matter disease on neuroimaging, with or without eye disease, in a patient from North America or Europe should strongly suggest the diagnosis of Baylisascaris NLM, and a history of exposure to raccoons or their feces should be sought. At present, in the absence of a brain biopsy, the diagnosis of B. procyonis NLM is dependent on serology. Demonstration of anti-B. procyonis antibodies in serum and CSF, particularly in the setting of a compatible clinical case and epidemiologic history, is the mainstay of diagnosis. Although basic laboratory tests, neuroimaging, and encephalography are, by themselves, nondiagnostic, they do provide important supportive evidence and may rule out other confounding diagnoses.
Anti-Baylisascaris antibodies can be demonstrated in CSF and serum by indirect immunofluorescence, enzyme-linked immunosorbent assay, and Western blotting. Enzyme-linked immunosorbent assay is the current test of choice. Serologic testing is currently available only from the Department of Veterinary Pathobiology at Purdue University, West Lafayette, IN (K. Kazacos). Indirect immunofluorescence tests use frozen sections of B. procyonis third-stage larvae as an antigen. Generally, in documented cases of NLM, good differential staining of larval sections is observed (41). Enzyme-linked immunosorbent assay and Western blotting use excretory-secretory products from in vitro cultures of B. procyonis larvae as the antigen (6). Larval excretory-secretory antigens have been characterized as complex glycoproteins, with molecular masses of 10 kDa to 200 kDa, that contain several different sugar residues (6). Protein epitopes of 33-kDa to 45-kDa antigens appear to be recognized selectively by antibodies from B. procyonis-infected humans and animals but not by normal human or T. canis antibody-positive sera (6). In addition, children with clinical B. procyonis NLM are strongly positive for anti-Baylisascaris antibodies in CSF and serum and have consistently been negative for anti-Toxocara antibodies (10, 16, 18, 29, 41, 43, 47, 51). In several of these cases, positive B. procyonis serology was confirmed by brain biopsy or at autopsy (8, 16, 25, 51). Acute- and convalescent-phase titers characteristically demonstrate severalfold increases in both serum and CSF anti-Baylisascaris antibody levels (18, 41).
In the absence of large population-based serologic studies, Baylisascaris seroprevalence is unknown. However, the finding of anti-Baylisascaris antibodies in asymptomatic family members of human cases and in individuals who have had contact with raccoons and preliminary results of a seroprevalence study in Chicago area children suggest that low-level asymptomatic infection may occur (9, 10) (W. B. Brinkman, K. R. Kazacos, P. J. Gavin, H. J. Binns, J. D. Robichaud, M. O'Gorman, and S. T. Shulman, Abstr. Pediatr. Acad. Soc. Ann. Mtg., abstr. 1872, 2003).
While a positive serologic test in a patient with compatible clinical signs and epidemiological risk factors is highly suggestive of the diagnosis, demonstration of larvae in tissues is confirmatory. Diagnosis of B. procyonis NLM has been made antemortem by demonstration of characteristic B. procyonis larvae in brain biopsies (8, 51). Nevertheless, the chances of isolating a portion of a larva in a small biopsy are low; Rowley and colleagues considered themselves fortunate to have made the diagnosis by biopsy in their case (51). If a diagnostic brain biopsy is considered, a deep white matter site is recommended. In addition to the fact that the diagnostic yield from a biopsy is potentially low, the availability of a serologic test has undoubtedly contributed to the decrease in numbers of diagnostic brain biopsies.
Although no routine laboratory test is considered diagnostic of B. procyonis NLM by itself, a number of studies provide additional supportive evidence. Most importantly, the presence of eosinophilia, particularly eosinophilic meningitis, should alert the physician to the possibility of a parasitic etiology (36, 49, 65). Eosinophilic meningitis, defined as the presence of 10 or more eosinophils/μl or eosinophilia of at least 10% in CSF, should never be ignored (36, 65). Eosinophils are not normally present in CSF; their presence narrows the differential diagnosis of CNS disease and provides an early or the only etiologic clue. In documented cases of NLM, the peripheral white blood cell count is usually mildly elevated (median, 17,400 cells/mm3; range 10,000 to 28,400 cells/mm3), but eosinophilia may be marked (median, 28%; range, 5% to 45%). Cerebrospinal fluid cell counts may be normal at presentation and generally demonstrate only mild leukocytosis (median, 16.5 cells/mm3; range, 1 to 125 cells/mm3), again with eosinophilia (median, 32%; range, 4% to 68%). Notably, even in the absence of pleocytosis, demonstrable CSF eosinophilia may be evident. Because eosinophils are easily missed in unstained or Gram-stained CSF, it may be necessary to request Wright's or Giemsa stain of cytocentrifuged CSF specimens. In documented cases of NLM, CSF protein is generally normal or only mildly elevated (median, 30 mg/dl; range, 18 to 69 mg/dl), while CSF glucose levels are normal (median 66 mg/dl; range, 41 to 81 mg/dl).
Although the finding of elevated serum isohemagglutinins, caused by cross-reactions between larval glycoproteins and human blood group antigens, is not specific for baylisascariasis, it does provide an additional clue to the diagnosis (6). Importantly, because B. procyonis does not complete its life cycle in humans, eggs or larvae are not shed in the feces of infected patients.
In acute B. procyonis NLM, changes on neuroimaging lag behind clinical disease and do not lead to earlier diagnosis. Initially, neuroimaging is frequently normal but eventually demonstrates only nonspecific diffuse white matter abnormalities, particularly of cerebral periventricular and deep cerebellar regions (Fig. (Fig.8)8) (10, 16, 18, 25, 41, 47, 51). This may be associated with changes of hydrocephalus, cortical edema, and loss of gray-white matter differentiation, with little or no meningeal or parenchymal enhancement. Evolution of clinical disease is manifest on neuroimaging as a progressive confluence of white matter changes and resolution of acute inflammation ending in residual global atrophy (18, 25, 41, 51).
Field studies have played an important part in the diagnostic workup of several of the documented cases of B. procyonis NLM, have contributed greatly to our understanding of the epidemiology and pathogenesis of the condition, and demonstrate targets for ongoing preventive efforts. In those cases in which formal field studies were carried out, unequivocal evidence of exposure to environments contaminated with raccoon feces has been demonstrated, and raccoon latrine densities have invariably been high (8, 10, 16, 18, 25, 47, 50). In Pacific Grove on the Monterey peninsula of California, Park and colleagues demonstrated B. procyonis eggs in fecal samples from 100% of latrines on a patient's property (21/21) compared to 44% (12/27) of fecal samples from latrines elsewhere in the city (47). Subsequently, a larger field study that included two additional northern California residential communities demonstrated a total of 244 latrines on 164 properties (50). Approximately half (44% to 53%) of the latrines contained B. procyonis eggs. Similarly, in southern California, a recent study of 800 raccoon latrines demonstrated 100% prevalence for B. procyonis eggs (14).
While peripheral and CSF eosinophilia, the hallmarks of B. procyonis NLM, are rare clinical entities, they occur under other infectious and noninfectious conditions (35, 65). Among helminth infections worldwide, Angiostrongylus cantonensis and Gnathostoma spinigerum are the most common causes of eosinophilic meningitis (36). Although angiostrongyliasis is endemic in southeast Asia, China, Japan, and the Pacific and Caribbean Islands, international travel increases the likelihood that such infections may be encountered locally (57). Angiostrongyliasis, in contrast to baylisascariasis, is generally associated with meningitis rather than meningoencephalitis, has a relatively benign course, and usually, but not always, carries a good prognosis. Gnathostomiasis, similar to baylisascariasis, may be associated with severe neurologic sequelae and a poor prognosis but is characterized by myeloencephalitis, focal cerebral hemorrhage (with xanthochromia), painful radiculopathy, and migrating cutaneous swellings. Neurocysticercosis, cerebral paragonimiasis, cerebral toxocariasis, neurotrichinosis, and CNS schistosomiasis are other, less common causes of eosinophilic meningoencephalitis. Although Toxocara spp. are a more common cause of VLM in humans, they are a very uncommon cause of NLM (65). Rare cases of cerebral toxocariasis are usually associated with signs of VLM and the presence of Toxocara antibodies in blood and CSF. Disseminated Coccidioides immitis infection, which is perhaps the most common cause of eosinophilic meningitis in the United States, is usually associated with intense basilar enhancement, hydrocephalus, acute infarction on neuroimaging, and positive CSF serology. Acute disseminated encephalomyelitis, which presents as an acute encephalopathy with cerebral white matter changes on neuroimaging in a previously healthy child, may be confused with B. procyonis NLM. However, acute disseminated encephalomyelitis is typically a monophasic nonprogressive illness, with more discrete multifocal gray and white matter abnormalities on neuroimaging, and generally has a good prognosis (51).
The prognosis for B. procyonis NLM is grave with or without treatment; among documented cases, there are no neurologically intact survivors. The majority of cases have been treated with anthelmintics and corticosteroids. Empirical anthelmintic treatment with thiabendazole, fenbendazole, tetramisole, or ivermectin has failed to prevent death or unfavorable outcomes. Anthelmintics successfully eradicate adult B. procyonis worms from the intestines of raccoons and skunks but are much less effective against larvae in tissues of intermediate hosts and humans (29). In one of the first-documented human cases, thiabendazole treatment failed to prevent the isolation of numerous live motile larvae from the child's brain postmortem (16).
Empirical treatment of suspected or proven NLM is extrapolated from studies of experimentally infected animals and individual case reports. Among currently available anthelmintics, animal data suggest that albendazole and diethylcarbazine have the best CSF penetration and larvicidal activity (29). Of the two, only albendazole has been used in children with NLM. In addition, albendazole appears to have the more favorable pharmacologic profile, with good absorption, high serum concentrations of the active metabolite, good penetration across the blood-brain barrier, and minimal toxicity (11, 26).
Experimental animal data and clinical observation of human cases emphasize the importance of the timing of treatment relative to larval invasion of the CNS. Anthelmintic treatment of experimentally infected mice was successful only if it was started before B. procyonis larvae entered the brain. Larvae enter the brain of mice as early as 3 days after ingestion, leading to signs of clinical infection by day 9 or 10 (29). However, laboratory mice fed B. procyonis eggs were protected from clinical NLM by treatment with albendazole (100%), diethylcarbazine (100%), mebendazole (80%), or thiabendazole (80%), daily from days 1 to 10 after infection (40). In contrast, when treatment started later, protection from clinical disease decreased with albendazole (75%) and diethylcarbazine (45%) treatment daily from days 7 to 10 after infection (29). In addition, it appears that treatment must be continued throughout the period of CNS vulnerability to larval invasion. Rates of protection from CNS disease were much lower in mice treated with albendazole (40%), mebendazole (20%), or thiabendazole (20%) from days 1 to 3 after infection only (40).
The majority of documented human cases were treated with thiabendazole (50 mg/kg of body weight/day) or, more recently, albendazole (20 to 40 mg/kg/day) for 1 to 4 weeks after presentation. However, because the diagnosis of B. procyonis NLM is considered only at the onset of clinical signs and symptoms, anthelmintic treatment has started late in the course of infection, at which stage larval invasion of the CNS has already occurred. Although ivermectin is effective in tissue nematode infections, such as lymphatic filariasis, its use should probably be avoided. It was unsuccessful in zoonotic B. procyonis NLM and did not appear to cross the blood-brain barrier (10, 29).
While treatment after the onset of symptoms has failed to prevent unfavorable outcomes, the role of prophylactic anthelmintic treatment for asymptomatic children is unclear but is felt to be potentially beneficial. Given the potentially devastating sequelae of untreated infection or late treatment and the availability of well-tolerated anthelmintics, prophylaxis appears warranted in select cases with documented exposure to infected raccoons, raccoon feces, or contaminated environments. Prophylactic albendazole has been started in children after exposure to raccoon latrines or cages, while results of tests to rule out contamination by B. procyonis are awaited (30, 44). In this situation, treatment has been discontinued after a few days when results of environmental tests were negative.
Systemic corticosteroids have proved beneficial as adjunctive anti-inflammatory treatment in neurocysticercosis and as single agents in angiostrongyliasis and ocular toxocariasis and have been used in the majority of B. procyonis NLM cases. In theory, corticosteroids are used to decrease the potentially deleterious eosinophilic inflammatory response and because of concerns that the larvicidal effect of anthelmintics may stimulate additional eosinophilic degranulation.
Baylisascaris procyonis OLM and DUSN have been treated successfully with the combination of laser photocoagulation and systemic corticosteroids to kill intraretinal larvae and decrease any resulting intraocular inflammatory response, respectively (21, 31). The role, if any, of anthelmintics in the treatment of B. procyonis OLM has not been established.
In the absence of effective treatment and early diagnosis, prevention of B. procyonis infection remains the best medicine. Most cases of B. procyonis infection are preventable by relatively simple measures. Education of the public regarding the potential dangers of contact with raccoons or their feces is the most important preventive step. Risk of infection is greatest when infants or toddlers with geophagia or pica come in contact with raccoon latrines or an environment contaminated by infected raccoon feces. Young infants and toddlers, particularly those with pica or geophagia, should be kept away from potentially contaminated areas. Parents should be observant for and discourage pica, and they should stress the importance of hand washing after outdoor play or contact with animals, including pet dogs. Raccoons should not be encouraged to visit homes or yards for food, and the keeping of raccoons as pets, particularly in households with young children, should be strongly discouraged. Downed timber should not be stored indoors or in areas readily accessible to inquisitive infants. Raccoon latrines in and around homes and play areas should be cleaned up and decontaminated. However, the longevity of B. procyonis eggs and their resistance to disinfection or decontamination makes successful environmental cleanup difficult. Detailed guidelines are available for raccoon latrine cleanup (29, 53). Heat is by far the best method of killing B. procyonis eggs (29). Boiling water, steam-cleaning, flaming, or fire are highly effective and practical methods for decontamination of large or small areas. The use of direct flames from a propane flame-gun is a favored method (29). For heavily contaminated areas a combination of removal and disposal of the top few inches of surface soil with flaming is most effective. Ideally, personnel cleaning contaminated areas should wear disposable overalls, gloves, and mask and eye protection. All potentially contaminated material removed from these sites, including used protective clothing, should be incinerated. Contaminated surfaces can be adequately cleaned with a xylene-ethanol mixture, after solid waste has been removed. However, chemical disinfection, in general, is rarely effective and not practical for large outdoor areas. Eggs are resistant to most common disinfectants; 20% bleach (1% sodium hypochlorite) will wash away sticky eggs but does not kill them (29).
While deworming wild raccoons may theoretically reduce the risk of infection in an area, interventions that combine raccoon depopulation and removal with cleanup and decontamination of latrines are likely to be more effective at reducing new and existing sources of infection (29, 44). If raccoons become a nuisance in residential areas, local authorities may be approached to trap and relocate or euthanize nuisance wildlife. However, if conditions in an area remain favorable for raccoons, such measures are unlikely to be effective in the long term. Simple and humane recommendations are available to protect humans and property while promoting harmonious coexistence with raccoons (66). Essentially, if there is no food or shelter to support them, most wild animals will go away.
Baylisascaris procyonis NLM is a potentially fatal, neurologically devastating infection, primarily of infants or young children. Although documented cases remain comparatively rare, the increase in populations of raccoons with B. procyonis infection living in close proximity to human residential populations suggests that the likelihood of human exposure and infection are high. While much has been learned concerning the epidemiology and pathogenesis of B. procyonis NLM, more remains to be elucidated. Large population-based seroprevalence studies should help to define the full clinical spectrum of infection and to identify at-risk populations most likely to benefit from targeted public health interventions. In the absence of effective treatment, prevention of infection remains paramount. Education of the public to the potential dangers is the most important first step.
We thank Tess Anton for assistance with access to the literature.