Abstract
In July 1998, the mother of an 18-month-old boy in rural Tennessee found a triatomine bug in his crib, which she saved because it resembled a bug shown on a television program about insects that prey on mammals. The gut contents of the Triatoma sanguisuga were found, by light microscopy and polymerase chain reaction (PCR), to be infected with Trypanosoma cruzi; PCR products hybridized with T. crazi-specific oligonucleotide probes. Whole-blood specimens obtained from the child in July and August were negative by buffy-coat examination and hemoculture but positive by PCR and DNA hybridization, suggesting that he had low-level parasitemia. Specimens obtained after treatment with benznidazole were negative. He did not develop and-T. cruzi antibody; 19 relatives and neighbors also were seronegative. Two of 3 raccoons trapped in the vicinity had positive hemocultures for T. cruzi. The child's case of T. cruzi infection—the fifth reported US autochthonous case—would have been missed without his mother's attentiveness and the availability of sensitive molecular techniques.
Chagas' disease (American trypanosomiasis) is endemic in parts of Latin America, where the causative agent, Trypanosoma cruzi, is transmitted among mammalian hosts by skin or mucous membrane contact with the feces of infected triatomine bugs [1]. In humans, the acute phase of infection, which lasts for weeks to months, often is asymptomatic but can be associated with mild nonspecific clinical manifestations (e.g., fever, lymphadenopathy) or with life-threatening myocarditis or meningoencephalitis. Identified cases of Chagas' disease should be treated with benznidazole or nifurtimox as early in the course of infection as possible; untreated persons are thought to be infected for life. The second phase of infection, the indeterminate phase, is asymptomatic. Years to decades later, 10%–30% of infected persons develop cardiac or gastrointestinal manifestations of chronic Chagas' disease.
During the acute phase, laboratory diagnosis typically relies on detection of circulating T. cruzi by microscopic examination of the buffy coat of a fresh blood specimen, hemoculture, or xenodiagnosis. Diagnosis during the indeterminate and chronic phases typically relies on serologic testing because parasitemia is difficult to detect with conventional parasitologic techniques. However, recent research has suggested that polymerase chain reaction (PCR), an investigational technique, is more sensitive and therefore may be useful for detecting low-level parasitemia and for monitoring response to therapy [2–6].
In the United States, an estimated 50,000–100,000 Latin American immigrants are infected with T. cruzi [1]. In addition, occasional US cases of Chagas' disease acquired through blood transfusion or vectorborne transmission are reported. Three of the 4 previously reported autochthonous vectorborne cases of which we are aware were in infants 2–3 weeks to 10 months of age in Texas in 1955 (2 cases) and 1983 [7–9]. The fourth case was in a 56-year-old woman in California in 1982 [10]. We describe here the fifth reported autochthonous case. This case occurred in Tennessee in 1998 in an 18-month-old child and was diagnosed by PCR.
Methods
Serologic testing, examination of buffy coat, and hemoculture
EDTA-anticoagulated human blood specimens were sent on ice packs to the Centers for Disease Control and Prevention (CDC). Plasma was removed and tested for anti-Tcruzi antibody by indirect fluorescent antibody (IFA) testing [11] and (for some specimens) by ELISA and latex agglutination (Wiener Lab, Rosario, Argentina). A wet mount and a Giemsa-stained slide of a drop of buffy coat were examined for T. cruzi trypomastigotes by light microscopy. Residual buffy coat and packed red blood cells were mixed together and either stored at −20°C until PCR was done or used for hemoculture as follows: Up to 1 mL of the mixture was placed in Novy-MacNeal-Nicolle culture medium with RPMI overlay, kept at 20°C-25°C for 6 weeks, and examined weekly for culture forms of T. cruzi.
Molecular analyses
The QIAamp DNA Extraction Kit (Qiagen, Valencia, CA) was used for extraction of DNA from 200 μL of blood or 100 μL of the intestinal contents (preserved in 5% formaldehyde) of the implicated bug. Extracted DNA was diluted 1 : 5, 1 : 50, and 1 : 500 in distilled water and stored at 4°C until use. PCR was conducted as described elsewhere [2, 3]. All reactions were done in 20-μL reaction volumes with 4 μL of diluted DNA per reaction. AmpliTaq Gold polymerase and PCR reagents were purchased (Perkin Elmer [Applied Biosystems], Foster City, CA). Amplifications were done in an integrated thermal cycler (ABI Prism 877; Perkin Elmer, Applied Biosystems) with the following sets of T. cruzi-specific primers: S-35 and S-36, which amplify a 330-bp minicircle sequence [2]; FPS-611 and FPS-761, which amplify a 177-bp sequence in a gene coding for flagellar protein [3]; and TCZ-1 and TCZ-2, which amplify 188 bp of a repetitive nuclear sequence [2, 9]. All positive PCR results were positive with at least 2 sets of primers. A β-actin primer set also was used (5′-GCT-GTGCTATGTTGCCCTAGAATTCGAGC-3′ and 5′-CGTACT-CCTGCTTGCTGATCCACATGTGC-3′) as a control. Reactions using the minicircle and TCZ primer sets were subjected to 30 cycles of amplification, versus 40 cycles for the FPS and β-actin primer sets.
All runs included additional controls. To control for possible inhibition of amplification by template DNA, additional PCR master mixes containing human DNA were spiked with 8 pg of DNA extracted from the quadriceps of T. cruzi-infected mice. The positive control consisted of 8 pg of this DNA but no human DNA, and negative controls had no template (distilled water) or DNA from uninfected human blood.
PCR products were electrophoresed in 2% agarose gels; visualized with ethidium bromide; transferred to nylon-supported nitrocellulose membrane; hybridized with specific probes for S-35/36 [2], FPS [3], TCZ [2], and β-actin (5′-GATGACCCAGATCATGTTT-GAGACCTTCAA-3′); and detected by chemiluminescence, as described elsewhere [12]. The CDC Biotechnology Core Facility synthesized the oligonucleotides used as primers and probes.
Animal and insect trapping
To determine whether local animals were infected with T. cruzi, traps were set nightly from 14 September to 17 September 1998, during hot, dry weather, on 4 nonworking farms that belonged to the patient's family and to neighbors on the same side of the road. Each farm was 4.07–8.09 hectares in size and abutted a heavily wooded area. On each farm, ∼50–150 Sherman-type traps for small rodents, baited with sunflower seeds or an oatmeal/peanut butter mixture, were placed along the fence lines, around buildings, and at other key sites (e.g., brush piles). Intermediate (squirrel)-sized Havahart traps, baited as above, were interspersed in much smaller numbers. About 12 large (raccoon)-sized Havahart traps, baited with canned sardines, were used per farm and were concentrated around outbuildings and more widely spaced along fence lines. Blood was obtained by cardiac puncture of small rodents euthanized with metafane. Larger animals were sedated with ketamine hydrochloride and released after femoral venipuncture. Buffy-coat examination and hemoculture were performed as described above. Attempts were also made to find triatomine bugs by searching in and around outbuildings and in materials on the ground (e.g., stacked wood, downed trees) and by hanging light traps in and around outbuildings.
Case Report and Investigation
The case
On 23 July 1998, the mother of a healthy 18-month-old boy in rural Rutherford County, Tennessee, found a bug in his crib and noted black spots resembling ink spots on his sheet. The sheet had last been changed ∼3 weeks earlier and had not been outside. She had not seen similar bugs in their 27-year-old brick house, which was in good condition and had woods on 3 sides. Because the bug resembled one she had seen on a television program about insects that prey on mammals, she asked an entomologist at a local university to examine it. On the basis of examinations there and at the CDC, the bug, an adult female Triatoma sanguisuga, was found to be engorged with blood and infected with T. cruzi. Motile trypanosomes were seen by phase-contrast microscopy of the bug's intestinal contents, and T. cruzi trypomastigotes and epimastigotes were seen by light microscopic examination of Giemsa-stained slides. T. cruzi DNA was detected by PCR with 3 sets of primers; PCR products of the appropriate size were visible on gels stained with ethidium bromide and hybridized with T. cruzi-specific probes.
In late July and early August, the child was febrile intermittently and had other nonspecific symptoms (table 1). His physical examination was unremarkable, except for pharyngeal erythema and multiple insect bite wounds of unknown type on his legs. His complete blood count was notable for mild anemia and lymphocytosis (table 1).
Clinical course and results of serial laboratory testing of a child in Tennessee who was infected with Trypanosoma cruzi.
Clinical course and results of serial laboratory testing of a child in Tennessee who was infected with Trypanosoma cruzi.
Multiple blood specimens were obtained from the child from July 1998 through March 1999, none of which had demonstrable anti-T. cruzi antibody or had T. cruzi visualized by buffy-coat examination or hemoculture (table 1). The first specimen tested by PCR and DNA hybridization was obtained on 30 July 1998; this and 2 subsequent pretreatment specimens repeatedly tested positive with at least 2 sets of primers and probes, and another laboratory also obtained positive PCR results (table 1). Therefore, the child was treated with benznidazole (Roche, Buenos Aires). He weighed 11.6 kg and received half a tablet (50 mg; 4.3 mg/kg body weight) twice daily from 17 August (25 days after discovery of the bug) through 19 October 1998. Except for mildly elevated aspartate aminotransferase values (table 1), he tolerated therapy well. Posttreatment blood specimens (October 1998 and March 1999) were negative by all testing methods.
Blood specimens from 19 other persons, 12 relatives (ages 3–56 years) who lived in or frequently visited the house where the child lived and 7 neighbors (ages 9 months to 61 years) from the farms where trapping was done, tested negative by microscopic examination, hemoculture, PCR (18 were tested), and IFA.
Animal and insect trapping
Seventeen wild or feral animals were trapped in the vicinity of the infected child's home: 8 rodents (4 white-footed mice, Peromyscus maniculatus; 2 wood rats, Neotoma floridana; 1 cotton rat, Sigmodon hispidus; and 1 shrew, Blarina cardinensis) and 9 larger animals (3 raccoons, Procyon lotor; 3 opossums, Didelphis virginiana; 2 skunks [blood from only 1 was cultured], Mephitis mephitis; and 1 feral cat, Felis domesticus). Specimens were also obtained from a dead skunk found nearby and from 2 outdoor dogs (Canis familiaris), ages 5–6 years and ∼1 year, that belonged to the patient's family. Two of 3 raccoons had hemocultures that, on the basis of morphologic criteria, were positive for T. cruzi. The shrew had a positive wet mount for a Trypanosoma lewisi-like organism, but the hemoculture was contaminated. Plasma from the older dog had an anti-T. cruzi IFA titer of 1 : 1024 that was confirmed by radioimmunoprecipitation assay (L. V. Kirchhoff, personal communication); hemoculture and PCR were negative. The dog was not examined but appeared to be healthy. One T. sanguisuga nymph found in a wood pile on a neighbor's farm was negative by microscopic examination and PCR of its gut contents. The family found 2 more adult female T. sanguisuga in June 1999, 1 in their basement and 1 on their front porch.
Discussion
In this investigation, we identified a T. cruzi-infected 18-month-old boy, the T. sanguisuga that transmitted the infection, 2 T. cruzi-infected raccoons, and a seropositive dog. To our knowledge, the child's case is only the fifth reported US case of autochthonous transmission of T. cruzi [7–10] and the first case in Tennessee. We attributed his case to vectorborne transmission because he had never had a blood transfusion, his mother had never traveled to Latin America and was seronegative for anti-T. cruzi antibody, and a T. cruzi-infected bug was found in his crib.
Although the bug's infection was confirmed both by conventional techniques that demonstrated the presence of the parasite and by investigational molecular techniques that demonstrated the presence of T. cruzi DNA sequences, our conclusion that the child was infected is based only on molecular evidence. Our evidence to support this conclusion is strong, although it would have been incontrovertible if the parasite itself had been detected (e.g., by hemoculture) or if serocon-version had been demonstrated. We consider it unlikely that the positive molecular results were false positives, because negative controls processed simultaneously with the patient's specimens were consistently negative and because of the strength of the data. Two laboratories obtained positive PCR results after separate DNA extractions; the CDC laboratory tested multiple replicates of multiple specimens with multiple T. cruzi-specific primer sets and probes. The fact that the primers and probes were for unrelated T. cruzi gene sequences suggests that the whole organism, rather than only DNA fragments, was present in the child's blood. The fact that PCR positivity persisted so long (table 1) suggests that the child was infected and did not simply have circulating, nonmultiplying T. cruzi from exposure to the bug's feces.
The disparity between the positive molecular results and the negative buffy-coat examinations and hemocultures probably is attributable to low-level parasitemia, which may have occurred in part because of the genetic constitution of the strain [2], and supports the conclusion of other investigators that PCR is a more sensitive diagnostic technique [2–6]. In a recent study of 59 Venezuelan cases of acute Chagas' disease, which were evaluated with conventional techniques typically 2–4 weeks after onset of symptoms, the sensitivities of xenodiagnosis, hemoculture, examination of Giemsa-stained blood smears, mouse inoculation, and microscopic examination of fresh blood were 61%, 53%, 34%, 29%, and 15%, respectively; 11 cases (19%) had negative results with all of these techniques, and 2 other cases were seronegative [13]. We did not attempt xenodiagnosis or mouse inoculation. The child's hemocultures may have been suboptimal because his young age precluded testing large volumes of blood, and transport time to the CDC caused delays in inoculating culture medium; of the 4 pretreatment specimens, the first 2 were inoculated into medium 4 days after they were obtained, and the latter 2 were inoculated 1 day after they were obtained (table 1). Possible explanations for lack of serocon-version, despite testing with 3 types of immunoassays, include institution of therapy relatively early in the course of infection and the presence of unidentified host or parasite factors.
Except perhaps for fever and mild anemia and lymphocytosis, which may have been from other causes, the child, like many infected persons, did not develop clinical manifestations of acute Chagas' disease. Although we do not know how virulent his T. cruzi strain was, he was treated with benznidazole to prevent later development of chronic Chagas' disease. PCR results for specimens obtained as late as 5 months after therapy were negative, which suggests that the infection was eliminated or the parasitemia was below the limit of detection.
The child's case of T. cruzi infection almost assuredly would have been missed if the bug had not been found, if his mother had not recalled watching a television program about insects that prey on mammals and had not asked an entomologist to examine the bug, and if molecular diagnostic methods that are both sensitive and specific had not been available. Given that infected triatomine bugs and mammalian hosts exist in the southern United States [14, 15], it should not be surprising that the sylvatic cycle occasionally spills over to infect a human. In addition, given that cases in humans with nonspecific clinical manifestations are easy to miss, it is likely that some other cases have been overlooked. However, autochthonous transmission is thought to be quite rare in the United States, in part because housing conditions generally are better than those in Latin America, thus limiting opportunities for contact between the vector and humans. We do not know the specific means by which the implicated bug (or the bug found in the basement in June 1999) got into or was inadvertently carried into the child's house, which was in good condition and did not appear to be bug infested. However, once the bug entered the child's crib, the bug's feces, which probably accounted for the black stains on the sheet, presumably had ample opportunity to get from the child's hands to the mucous membranes of his mouth or conjunctivae or to be rubbed into the insect bite wounds of unknown type on his legs. The fact that 4 of the 5 reported autochthonous cases have been in children<2 years old suggests that young children have behaviors that increase their risk for exposure to triatomine feces.
Acknowledgments
We thank Raymond F. Beach, C. Ben Beard, J. L. Butcher, Jerry Franklin, Carter Garner, Dennis Juranek, L. Vaughn Kirchhoff, Fernando Mornay-Linares, Gordon Moreau, Pamela Pennington, Norman Pieniazek, and Melinda Tibbals for contributions to the investigation.
