Characteristics of both urban and rural populations are shown in Table 1. Results of serologic testing for 76 (69%) and 38 (44%) persons from urban and rural populations, respectively, were positive for T. cruzi infection. Globally, 80 (41%) persons did not fulfill criteria of serologic diagnosis of Chagas disease (in all cases, IFA test was negative). In eight of nine rural seronegative patients, anti-cruzipain antibodies were investigated with negative results.

T. cruzi was detected by PCR amplification of a nuclear DNA fragment by using the O1/O2 primers (see Materials and Methods). This reaction has been previously demonstrated to be highly specific to detect T. cruzi in blood samples (12). Parasitemia by PCR assay was detected in 34 (17%) of 194 persons and was more frequently found in rural than in urban populations (20 and 14 positive persons, respectively; p = 0.05) (Figure). When only the seronegative population was considered, PCR was positive in 12 (15%) persons (3 and 9 from urban and rural population, respectively; p = 0.36). Clinical characteristics of these patients are shown in Table 2. Only one patient (from the urban group) had a previous positive Machado Guerreiro test. Two of three urban patients were born in a highly disease-endemic area. Disease in all of these urban patients was controlled a year after recruitment, and subsequent serologic testing was negative. Of the rural case-patients (born and living in La Posta), none reported previous positive serologic findings. EKG and echocardiogram were performed for four patients from a rural area (Table 2).

When T. cruzi infects a mammal, several immunologic reactions occur that eliminate the parasite. First, a cellular immune response attempts to isolate the microorganism and avoid its wide spread. Simultaneously, a humoral response occurs, with IgM antibodies first and IgG antibodies 2–3 weeks later. However, because of the lack of efficacy of these mechanisms, the parasite persists in low-density tissues and in turn, triggers an inflammatory response, resulting in tissue damage during the chronic period of the disease (14). Parasites are rarely isolated from blood or tissue from chronically infected patients, and the diagnosis is based on serologic analysis.

In our study, we observed that persons with positive T. cruzi in blood and negative serologic findings could be detected in a population with high epidemiologic risk. This observation has been previously reported in Wincker et al. (5), who studied PCR technique using serum samples from 45 Bolivian children. They found two positive PCR results in 17 seronegative children, and in one of them, parasites were seen on direct blood examination. These authors also reported a patient with the same infectious condition in 268 children with high epidemiologic risk for Chagas disease (6). In Brazil, Avila et al. (7) observed three discordant cases, one of which had typical findings of myocardial damage. Similarly, Castro et al. (9) detected 3 persons with positive PCR results among 9 seronegative controls, and Gomez et al. (8), reported 10 positive PCR results in seronegative patients of 110 residents of a highly disease-endemic region.

Several arguments have been proposed to explain this controversial situation. Recent infection that has not yet been recognized by the immune system of persons highly exposed to vectorial infections is one possible explanation. However, acute infection is not a frequent event in our study population because of age and because urban population is infrequently exposed to vectorial reinfection. Alternatively, one could speculate that positive samples could have been contaminated with DNA, but this theory has been disregarded by many authors (8,9). We repeated serologic and PCR assays three times for each patient with two different operators, and we obtained the same results. Finally, T. cruzi may chronically infect some patients but a humoral response may not develop or be detected by conventional serologic testing. Addressing this point, Castro et al. (9) observed that 80% of seronegative but positive PCR patients had lytic antibodies against T. cruzi by a complement-mediated lysis test (CoML). Similarly, Leguizamón et al. (15) have reported patients who were seropositive for Chagas disease only by inhibition transamidase assay but negative with conventional serologic testing. To test this hypothesis, we searched for anti-cruzipain antibodies in eight rural patients, but all of them were negative.

Otherwise, independently of its cause, considering the clinical and diagnostic consequences of this phenomenon is necessary. In our study, at least 3 of the 12 patients with high epidemiologic risk for Chagas disease had signs of cardiac compromise. The consensus is that the detection of DNA constitutes real proof of parasites. DNA detected in blood is originated from extracellular parasites recently liberated or destroyed. According to this theory, Tarleton and Zhang observed that after injection of high doses of kinetoplastic DNA (kDNA) of T. cruzi in muscle, kDNA is detected in blood 2 days later (16). However, since the parasite is infective as tripomastigotes but not as a portion of DNA, we cannot be sure that DNA detected by PCR in blood is a reliable surrogate of infecting T. cruzi forms. One alternative is the possibility that fragments of amastigotes (the tissue-infecting forms) reach the bloodstream after interacting with the immune system. Even though this consideration may be found relevant in the design of transfusion policies, no reliable information currently exists, and the potential for these persons to transmit the disease is still uncertain.

Comparing our results with a population without any risk for Chagas disease to determine the predictive value of PCR among seronegative persons would be interesting. However, the accuracy of PCR has been studied in depth (2–4,12,17). While previous reports of PCR in Chagas disease used a sequence of kDNA of T. cruzi to detect the parasite, we used nuclear DNA that has been also validated (12,13). Currently, we have not carried out a systematic comparison of the PCR sensitivity for different T. cruzi sequences. We considered determining which of the different PCR systems cited in the literature is the more sensitive and accurate for detection of parasitemia in blood specimens of patients with chronic Chagas disease.

In summary, we found a prevalence of 15% of T. cruzi DNA for American trypanosomiasis in a seronegative population living in Chagas-endemic regions. We also observed that some of these persons had cardiac abnormalities suggestive of Chagas cardiomyopathy. Experts should consider these finding when making diagnostic, therapeutic, and transfusion recommendations.