Of the 1,973 persons registered during the census, 137 (7%) were <5 years of age, which left 1,836 potential study participants. Of these, 949 persons (51.7%) were examined. All 949 were evaluated with ultrasonography, 829 had chest radiographs taken, and 929 contributed blood samples. The proportion of females was higher among participants (60.3%) than nonparticipants (50%) (p<0.05). Of those with ultrasonographic results, 39 had no serologic results because they refused to have their blood collected. In addition, 125 persons had serologic results but refused to have chest radiographs taken. The ages of participants and community members were similar (mean age 28.8 and 28 years, respectively) when children <5 years of age were excluded from analysis.

Ultrasonography showed prevalence of CE in the liver to be 4.7% (45/949); radiography showed prevalence of CE in the chest to be 1.1% (9/829). Two persons had cysts in the liver and lungs. Therefore, the CE prevalence according to ultrasonography and radiography was 5.5% (52/949; 95% confidence interval [CI] 4.1%–7.1%). Seropositivity according to IBCF was 8.9% (83/929; 95% CI 7.2%–10.9%); seropositivity according to rEpC1-GST was significantly higher at 19.7% (184/929; 95% CI 17.2%–22.4%, p<0.01). All 83 IBCF-positive persons reacted to the 16-kDa band; the bands of 21 and 8 kDa were observed for 78 and 76 persons, respectively. No differences were found among the 9 communities (Table 1). When participants were divided into 2 groups according to median age (<23 and >23 years), no difference was found for proportion of those who were positive by serologic assay and by radiographic examination; however, a significant difference existed according to ultrasonographic examination (14/456 [3.1%] for those <23 years vs. 31/493 [6.3%] for those >23 years; p<0.05).

The total number of liver cysts was 50 (5 persons had 2 liver cysts each); the total number of lung cysts was 10 in 9 persons. The lung-to-liver ratio was 1:5. Most of the liver cysts were classified as CE5 (54%, 27/50), inactive cysts with calcified walls, followed by CE1 (20%, 10/50) active cysts and CE2 (10%, 5/50). Types CE3 and CE4 with signs of initial degeneration (8% each, 4/50) were rare (Figure 2). The average age was similar for persons with CE1, CE4, and CE5 at 40.6, 40.9, and 40 years of age, respectively. Those with CE2 and CE3 averaged 28.8 and 23.3 years of age, respectively.

The frequency of persons in the group that had CE-positive imaging results who also had IBCF-positive results was 47.1% (8/17; 95% CI 23%–72.2%) for hepatic noncalcified cysts, 35.7% (15/42; 95% CI 21.6%–52%) when hepatic calcified cysts were included, and 22.2% (2/9; 95% CI 2.8%–60%) for pulmonary cysts. Of the 52 persons who had CE in liver, lungs, or both, 49 provided serum samples (3 who had positive ultrasonography results did not provide a blood sample). The frequency of persons in the group that had CE-positive imaging results who also had rEpC1-GST–positive results was 16.7% (7/42; 95% CI 7%–31.3%) for all hepatic cysts, and 29.4% (5/17; 95% CI 10.3%–56%) when hepatic calcified cysts were excluded. This test detected 33.3% (3/9; 95% CI 7.5%–70%) of those with CE-positive chest radiographs. Neither serologic test detected >50% of persons with presumptive CE imaging results (Table 2). However, the IBCF was >2× as sensitive as the rEpC1 for detecting persons with CE-positive imaging results.

Of the 2 persons who had cysts in lung and liver, 1 had positive results for both serologic tests and the other had negative results for both. The frequency of seropositivity for persons with liver cysts <20 mm in diameter was 25% (6/24) for IBCF and 4% (1/24) for rEpC1-GST. In contrast, the frequency for persons with liver cysts >20 mm in diameter was 50% (9/18) for IBCF and 33% (6/18) for rEpC1-GST. A significant difference in detecting cysts with diameters <20 mm versus >20 mm was found for only rEpC1-GST (4%– 33%, p<0.05).

The agreement between IBCF and rEpC1-GST was only 8%, which would be expected by chance alone (κ = 0.08, p<0.01). Of 928 participants, 26 were positive by both serologic tests (2.8%), and 688 (74.1%) were negative by both. A total of 157 persons who had IBCF-negative results had rEpC1-GST–positive results; 57 who had IBCF-positive results had rEpC1-GST–negative results. CE-test positivity was not significantly associated with sex according to any diagnostic test (Table 3).

To assess possible cross-reactions with cysticercosis, we evaluated serum by using a purified glycoprotein Taenia solium antigen, in an immunoblot format (18). At the population level, antibodies to T. solium cysticercosis did not affect hydatid serologic assay results. The seroprevalence of cysticercosis was statistically similar for hydatid-seropositive and hydatid-seronegative persons, regardless of the antigen used to diagnose hydatidosis (IBCF: 11/83, 13.3% vs. 103/852, 12.1%, respectively; rEpC1-GST: 24/184, 13% vs. 90/750, 12%, respectively). Similarly, at the population level, the seroprevalence of E. granulosus antibodies was similar for cysticercosis-seropositive and cysticercosis-seronegative persons, regardless of the antigen used to detect Echinococcus (IBCF: 11/114, 9.7% vs. 72/821, 8.8%, respectively; rEpC1-GST: 24/114, 21.1% vs. 160/820, 19.5%, respectively).

This study demonstrates and confirms the high prevalence of CE in humans in the central Peruvian Highlands. It also highlights the limited performance of 2 immunoblot tests (IBCF and rEpC1-GST) under field conditions by detecting <50% of persons who had CE-positive imaging results. We show the utility of ultrasonography for CE screening, which demonstrated an elevated percentage (54%) of apparently inactive and calcified hepatic cysts.

The survey was well accepted by the study population. However, most of the participants were women; men tend to be more reluctant to participate in medical studies, especially those involving blood sampling. The CE prevalence of 5.5% found by ultrasonography and radiography is similar to previously reported rates from other central highland communities in Peru (4.9%–5.7%) (12,14). However, a study in Vichaycocha (north highland) showed a rate of 9.3% by ultrasonography and radiography, and a seropositivity rate of 18.2% (11). These rates are comparable to the highest reported prevalences in other countries such as China (liver CE from 3.3% to 6.6% [22,23]), Kenya (5.6% liver CE [1]), and Argentina before initiation of its control program (5.6% in school children [24]).

Unlike some other areas of Peru, Yanahuanca has not had a CE control program. This might explain, at least in part, why the overall prevalence in our survey was as high as 12.5% when ultrasonography, radiography, and IBCF results were combined and up to 23.8% when ultrasonography, radiography, and rEpC1-GST results were combined. Notwithstanding, these figures do not represent the true disease prevalence because they might reflect the continuous transmission and endemicity of E. granulosus in this region. Diagnostic approaches for CE based on imaging techniques can be problematic because of variations in size, shape, and location of the cysts. In addition, E. granulosus distribution, host susceptibility, and strain variation might affect disease transmission in different areas of the Peruvian Highlands (to our knowledge, no studies have tried to characterize E. granulosus strains in Peru).

One of the underlying weaknesses of this study was the lack of a true standard (a test with 100% sensitivity and 100% specificity), which would enable evaluation of alternative diagnostic tests and underlying prevalence. Most areas of medicine lack a true standard, yet recent statistical techniques have been developed that can help evaluate diagnostic tests and estimate true prevalence in the absence of such a standard. Most of these techniques rely upon a Bayesian framework (25–27) and are computationally intensive but more flexible than maximum likelihood–based approaches because they can incorporate correlation among diagnostic tests. Although the Bayesian approach offers distinct advantages, potential problems include specification of an appropriate prior distribution and a nonidentifiable model. Our study encountered both of these problems because reliable prior information is not readily available and estimates provided by a Bayesian approach are limited by the lack of identifiable groups.

The poor performances of the IBCF and rEpC1-GST testing may be related to false-positive imaging results from other space-occupying lesions (e.g., neoplasia, abscesses, nonparasitic cysts). Additionally, participants might have had low or undetectable levels of circulating antibodies from different stages of cyst development or degeneration. The production of IgG depends on the number, size, location, and condition of the cysts; only 60%–80% of persons with confirmed CE become seropositive (28). Calcified cysts are less seroreactive, thus decreasing seropositivity (16,22), as observed in this study. Previous studies using the same IBCF testing found that the frequency of seropositivity was 57% (12) and 53% (14) for liver hydatid cysts (ultrasonography), and 13% for lung cysts (11), similar to what we found in this study. However, another study, in which most of the liver cysts were active, reported a proportion of persons who were positive according to IBCF testing to be as high as 73% (11).

Other possible reasons for the limitation of serologic testing might be the weak immune response against pulmonary cysts, cysts at other sites (e.g., brain, eyes, bones, ovaries), small or poorly defined cysts, and a thick collagen cyst wall that would reduce antigen exposure (29). However, increased seropositivity (up to 50%) for detecting large hepatic cysts (>20 mm in diameter) has not been reported previously and may be due to elevated antigen concentrations in these cysts. To our knowledge, the only study showing a correlation between cyst size and seropositivity was performed in sheep (30).

The different antigen sources would explain the disparity and poor agreement between the 2 serologic tests. EpC1 is a recombinant antigen obtained from protoscolex larvae from sheep hydatid cysts (17), while the IBCF uses bovine hydatid cyst fluid (16). The IBCF appears to be more responsive than the rEpC1-GST in detecting CE-image cases. Crude hydatid fluid has been recommended for mass serologic screening (31) and purified antigen 5 (Ag5) and AgB for specific diagnosis. Ag5 and AgB are recognized as 2 of the most useful E. granulosus antigens for diagnosis (32), although Lorenzo et al. (33) found that hydatid cyst fluid, AgB, and its subunit AgB8/1 exhibited equivalent diagnostic efficiencies in a randomized multicenter study.

Among participants with CE-negative ultrasonography images, rEpC1-GST testing detected >3× more seropositive persons than the IBCF (19% vs. 7.3%, respectively). This scenario has been described with other diagnostic test such as AgB ELISA, which detected 5.3% CE-seropositive persons in a group with CE-negative ultrasonography images (22). We do not have evidence of cross-reaction with Fasciola for any of the serologic tests used in this study; however, Hymenolepis nana, Entamoeba histolytica, Giardia lamblia, and Taenia sp (34) are endemic to the study area, which might affect the serodiagnosis. We demonstrated no cross-reaction with cysticercosis (T. solium) by IBCF or rEpC1-GST because the proportion of persons who were cysticercosis positive was equal among those who were CE seropositive and CE seronegative according to both assays. Other possible explanations include past exposure to Echinococcus eggs (aborted infection) that produced only transient antibodies (35), and undetected cysts.

In our study, the lung-to-liver ratio of 1:5 is higher than ratios reported in other epidemiologic studies in Peru (11) but within the range of those reported in other Echinococcus-endemic South American countries such as Argentina, Chile, and Uruguay (1:3–1:13) (11,36). Molecular genetic studies of E. granulosus in the Peruvian Highlands may clarify some issues about the organ infection preferences (tropism), tissue survival, infection rates, immune responses, and the performance and agreement of immunodiagnostic tests.

Among liver cysts, 54% were CE5 and 20% were CE1; other studies have typically displayed an exponential decline in the frequency of liver cyst types from CE1 (most frequent) to CE5 (most rare) (37). Because most of these persons had not received a diagnosis of hydatid disease or antihelminthic treatment, these calcified forms are most probably the result of the natural process of degeneration driven by individual immune responses (38). A proportion of hydatid cysts die after initial establishment; thus, calcified lesions can be observed macroscopically (29). The geographic variation of cyst type frequencies can also depend on the time between infection and evaluation, the immune response of a highly exposed population, and E. granulosus genetic variation (15). Calcified cysts may also have been misdiagnosed with other lesions from biliary cysts, pyogenic abscesses, amebic liver abscesses, or even tumor-like masses or metastases (39), although these conditions are uncommon in Peru. We did not study the specific IgG subclasses in relation to cyst types, but their quantification may be important for understanding the natural history of hydatid cyst. IgG₄ antibody response is associated with development, growth, and progression (CE1 to CE3); IgG₁, IgG₂, and IgG₃ occur predominantly when cysts became infiltrated or degenerated (CE4 and CE5) (40).

Our study provides data on CE in the surveyed communities and shows the results of using ultrasonography, radiography, and immunodiagnosis for large-scale population screening. Determining baseline prevalence with ultrasonography enables the evaluation of epidemiologic surveillance activities and study of the natural history of CE. Ultrasonography is well accepted by the population and is relatively less expensive than other imaging techniques. Methods that are inexpensive and relatively easy to use, such as immunodiagnosis and ultrasonography, are required for large-scale screening of populations in which hydatidosis is endemic. However, serologic assays have serious limitations under field conditions, as has been demonstrated in this study. Seroepidemiologic surveys for CE require better diagnostic antigens and should be supported by imaging methods whenever possible.