Epidemiology of Echinococcus granulosus infection in the central Peruvian Andes.

Epidemiology of Echinococcus granulosus infection in the central Peruvian Andes P.L. Moro,1 J. McDonald,2 R.H. Gilman,3 B. Silva,4 M. Verastegui,5 V. Malqui,6 G. Lescano,7 N. Falcon,8 G. Montes,9 & H. Bazalar'° The prevalence of human, canine, and ovine echinococcosis was determined in an endemic area of the Peruvian Andes where control programmes have not been operational since 1980. Prevalence of infection in humans was determined using portable ultrasound, chest X-rays, and an enzyme-linked immuno- electrotransfer blot (EITB) assay. Canine and ovine echinococcal prevalence was determined by micro- scopic stool examinations following arecoline purging for tapeworm detection and by examination of the viscera from slaughtered livestock animals, respectively. The prevalence among 407 humans surveyed was 9. 1%. The frequency of disease in the liver, lung, and in both organs was 3.4%, 2.0%, and 0.2%, respectively. Portable ultrasound or portable chest X-ray has shown that, compared to adults, children under 11 years had significantly higher seropositive rates without evidence of hydatid disease (P < 0.05). Among the 104 dogs inspected for echinococcus after arecoline purging, 33 (32%) were positive for adult tapeworms. Among the 1 17 sheep slaughtered at the local abattoir, 102 (87%) had hydatid cysts. The prevalence of human hydatidosis in this endemic area of Peru is one of the highest in the world and nearly five times higher than previously reported in 1980. An increase in echinococcosis prevalence may result after premature cessation of control programmes. Introduction Human hydatidosis caused by Echinococcus granu- losus is endemic in the sheep-rearing regions of Peru (1, 2), where echinococcosis has been recognized as a major public health problem (1-3). However, preva- lence surveys have not been conducted in intermedi- ate, definitive, or accidental hosts since 1980; most I Research Associate, Asociaci6n Ben6fica PRISMA and the Department of Pathology, Universidad Peruana Cayetano Heredia (UPCH), Lima, Peru. 2 Instructor, Department of Internal Medicine, University of Utah, Salt Lake City, UT, USA. 3Professor, Department of International Health, The Johns Hopkins School of Hygiene and Public Health, 615 Wolfe St., Baltimore, MD 21205, USA. Requests for reprints should be sent to this author at this address. 4Staff Physician, Department of Radiology, Hospital Nacional Arzobispo Loayza, Lima, Peru. 5Adjunct Professor, Department of Pathology and Auxiliary Pro- fessor, Department of Microbiology, UPCH, Lima Peru. 6 Research Associate, Department of Pathology, UPCH, Lima, Peru. 7Project Coordinator, Asociaci6n Benefica PRISMA, Lima, Peru. 8 Research Associate, Public Health Section, Faculty of Veterinary Medicine, Universidad Nacional Mayor de San Marcos (UNMSM), Lima, Peru. 9 Chief, Veterinary Services, Tupac Amaru, Junin, Peru. 10 Professor, Faculty of Veterinary Medicine, UNMSM, Lima, Peru. Reprint No. 5812 studies were carried out in the 1970s in the central Andean department of Junin (4). At that time a pilot hydatidosis control programme was initiated in the farming cooperative Tupac Amaru (3), but these efforts were abandoned 17 years ago largely due to political instability. In 1980, when the control programme ceased, the prevalence of echinococcal infection in dogs and sheep was 1.6% and 5%, re- spectively (3, 5). Use of enzyme-linked immunolectrotransfer blot (EITB) and double-diffusion (DD5) assays on serum collected in 1989 from Tupac Amaru revealed a seroprevalence of 1.9% in 309 individuals (6). While the EITB assay is useful in the diagnosis of human hydatid disease, the sensitivity is only 65%; hence the prevalence was probably an underestimate (7). The present epidemiological study was con- ducted in Tupac Amaru from September 1994 until October 1995, using the EITB assay, as well as portable ultrasonography (US) and portable chest X-rays to determine the prevalence of human hydatidosis. A case-control study was undertaken to identify environmental and cultural factors impor- tant in E. granulosus transmission, and the preva- lence of echinocococcal infection in farm animals and dogs was also determined. This study was approved by the ethical review boards of the Universidad Peruana Cayetano Heredia, the Universidad Nacional Mayor de San Bulletin of the World Health Organization, 1997, 75 (6): 553-561 © World Health Organization 1997 553 P.L. Moro et al. Marcos, The Johns Hopkins University, and the Council of Tupac Amaru. Free and informed con- sent was obtained from the study subjects. Materials and methods Study site. Tupac Amaru is a farming cooperative occupying an area of 21.7 km2 in the central Peruvian Andes at altitudes between 3600 m and 4300 m above sea level. The terrain is mostly mountainous with few valleys and streams. Sheep are the domi- nant livestock (99 300 animals), but cattle (3370) and alpacas (4300) are also raised for human consump- tion; horses (609) are used as pack animals. Pigs are absent, except in nearby rural communities not be- longing to Tupac Amaru. The cooperative consists mainly of adobe houses with wooden floors and piped river water, but most have no electricity. Families use toilet drains, borehole latrines or nearby streams for defecation. Some workers and their families live in these houses, while the shepherds live in widely dispersed shacks with dirt floors and no electricity. Shepherds rou- tinely use dogs to care for the flocks of cooperatively owned sheep. Medical facilities are often 5-20 miles distant. Study design. A community census was performed first, the questions to families including those on their knowledge of hydatidosis transmission. The study was carried out in two phases. First, surveys of humans, dogs, and livestock for echinococcal infec- tion; second, a case-control study to determine fac- tors important in E. granulosus transmission. Epidemiological survey. Meetings were held to ex- plain the purpose of the study to members of the cooperative. Two teams conducted the survey. A medical team carried out ultrasonic scanning of hu- man subjects older than 5 years and collected serum samples. A second team of veterinarians purged the dogs and collected faecal samples, and examined the viscera of slaughtered livestock for echinococcal cysts. The teams worked on market days when the shepherds frequently come to the cooperative. Cases with presumptive hydatidosis were in- formed of the findings and referred by the physician in the health post to a national health insurance hos- pital in Oroya or Huancayo for surgical or medical treatment. Ultrasound survey. Individuals were examined in the supine position, using a Shimasonic (Shimadzu, Nishinokyo-Kuwauaracho, Kyoto, Japan) SDL-32, 3.5 MHz, portable ultrasonograph powered by a portable 4-kW generator, with liquid paraffin as the transducing medium. Cysts were diagnosed using the scheme described by Gharbi et al. (8). Fluid-filled hepatic cysts without daughter cysts (solitary) and fluid-filled cysts with septa (daughter cysts) were considered to be hydatid cysts. Hepatic cysts with a heterogeneous echographic pattern or calcification were considered to be presumptive hydatid cysts (8, 9). Renal cysts were not considered as hydatid, since idiopathic renal cysts are common (10). Enzyme-linked immunoelectrotransfer blot (EITB) assay. An EITB assay for human hydatid disease diagnosis, which identifies three antigens of M, = 8 x 103, 16 x 103 and 21 X 103, was performed as previ- ously described except for the following modifica- tions: antigen was used at twice the previous concentration (0.2 mg/ml) and 20 ,l of human serum at 1: 50 dilution was used instead of 10 gl at 1: 100 dilution (7). A positive result was defined as the presence of any of the above bands. Chest X-ray. Posterior-anterior chest radiographs were taken at the cooperative's health post in March 1995 using a portable machine. Films were read by a radiologist (BS), with no knowledge of the serological and ultrasound results. Technically defi- cient chest-X rays were repeated. Beggs' criteria for lung hydatid disease were used for diagnosis of pul- monary hydatidosis (11). Arecoline purging of dogs. Dogs were given an oral dose of arecoline hydrobromide (4 mg/kg), as previ- ously described (12). Faecal samples were collected from dogs that purged successfully according to WHO guidelines (12). Slaughterhouse visits. The local abattoir was visited on slaughter days and the percentage of infected animals with hydatid cysts was noted. Conditions in the abattoir such as veterinary supervision, facilities for infected viscera disposal, and presence of dogs were observed. A total of 117 sheep and 25 cattle were inspected during a 2-month period. Pigs were not killed in the abattoir, but those in a community one kilometre away were inspected. Surgical incidence of hydatidosis. Individuals from the cooperative with hydatid disease who were not part of our screening study were located by review- ing the local health post's records. All the workers were covered by national health insurance, but must first be evaluated at the local health post before re- ferral to one of the national hospitals for surgery or other treatment. WHO Bulletin OMS. Vol 75 1997554 Epidemiology of Echinococcus granulosus infection in Peru Data analysis. Individuals who were examined by EITB, ultrasound, and chest X-ray were designated the examination cohort (EC). Those not surveyed at all or surveyed by only one or two methods were not part of this cohort and were designated the non- examination cohort (non-EC). The x2 test was used to determine differences in categorical variables between the EC and the non- EC groups. Kappa analysis was used to test agree- ment between the different diagnostic methods used among EC subjects. Case-control study design. Infected study subjects were compared to noninfected cooperative resi- dents to identify the risk factors for hydatid disease. Cases of E. granulosus infection were defined as individuals with a positive EITB assay, and/or those with presumed liver or lung hydatid cysts diag- nosed by ultrasound scanning or chest X-rays. Con- trols were defined as individuals who participated in the study and had a negative EITB, ultrasound and chest X-ray findings and no previous history of hydatid disease. One case was matched for age (±5 years) and sex with two non-infected con- trols. Controls were selected from the list of parti- cipants who met the above criteria using a random table. Cases and controls had a lateral chest X-ray taken at a clinic in Huancayo city to rule out hid- den cysts, in addition to the posteroanterior chest X-ray. Exposure of cases and controls to E. granulosus was assessed using a questionnaire to identify prac- tices which might increase such exposure and knowl- edge regarding transmission. Questions included place of birth, level of education, and occupation; type of farm animals raised and/or slaughtered, and the place of slaughter (home, field, abattoir) and disposal of infected viscera; the number of dogs kept and their use (shepherd or pet), the place where they slept, and the type of food given to them. The individual's knowledge of what caused hydatidosis, how it was transmitted, and what type of disease it caused in humans and farm animals was also determined. Data were analysed by standard methods for matched case-control studies. For the questionnaire data, odds ratios (OR) with 95% confidence limits were calculated for each item in the questionnaire. Questionnaire items that were significant at P < 0.05, or with high OR but borderline statistical sig- nificance (P < 0.2) in the initial univariate matched- pair analysis, were tested for inclusion in the logistic models. The conditional logistic statistical model for matched data implemented by Statistics Graphics Data Management (Stata Statistical Software: Re- lease 4.0. Stata Corporation, College Station, TX, USA, 1995) was used to identify questionnaire items that were independently associated with an in- creased risk of E. granulosus infection. Results Human study The census revealed 1042 individuals living in 238 households (52% males; mean age, 23 years (range 0-72 years)). As 155 (15%) were <6 years, there were 887 subjects who were potentially eligible for the study. The census revealed that 82% of house- holds admitted they had seen hydatid cysts in ani- mals, 65% had complete or partial knowledge of the cause of hydatid disease, 93% thought it was danger- ous to both humans and animals, and 84% regarded dogs as a cause of transmission of the disease to humans. A total of 407 individuals in the examination cohort (EC) were surveyed by all three methods. The EC and non-EC individuals were similar for most of the characteristics tested (Table 1). Defining hydatid infection as a positive result in at least one of the three tests (EITB, ultrasound, and chest X-ray), we found the prevalence of human hydatidosis to be 9.1% (Table 2). Hydatid disease, defined as a positive result by ultrasound and/or chest X-ray (8, 11), was present in 5.7% of individu- als (23/407). The ratio of lung to liver cysts was 1:1.75. The frequency of seropositivity was 57% (8/14) in those with presumptive liver hydatid cysts and 38% (3/8) in those with presumptive lung hydatid cysts; the only case with cysts in both lung and liver was also EITB-positive. The overall sero- positivity of presumptive images in any one organ was 52% (12/23). Fourteen individuals were seropositive but were negative by ultrasound or chest X-ray (Table 3). None of the 23 individuals in whom presumptive liver or lung cysts were detected had symptoms associated with these lesions and only three had hepatomegaly on physical examination (Table 4). Four individuals with EITB-negative renal cysts lacked ultrasound criteria for hydatid cysts and were excluded from disease prevalence. Age, sex, and the specific findings by EITB, ultrasound and chest X- ray of the 23 individuals with presumptive hydatid cysts are shown in Table 4. About 50% of family members were examined by all three methods (407/810). Subjects positive to any one method belonged to 34 households with 186 family members, 52% of whom (97/186) had three examinations performed. Three households had two members each with hydatid infection. WHO Bulletin OMS. Vol 75 1997 555 P.L. Moro et al. Table 1: Characteristics of 407 individuals surveyed by three methodsa for hydatid disease and of those not surveyed In Tupac Amaru, Peru, 1994-95 Examination Non-examination Odds 95% confidence cohort cohortb P valuec ratio interval Age group: 6-10 years 85 (20.9)d 87 (18) NSe NS 11-15 years 74 (18.2) 101 (21) NS 16-30 years 71 (17.4) 138 (29) <0.001 0.52 0.37-0.73 31-40 years 82 (20.1) 56 (12) <0.001 1.91 1.30-2.82 41-50 years 54 (13.3) 45 (9) NS >50 years 41 (10.1) 53(11) NS Total 407 (46) 480 (54) Mean age (years) 27 26 NS NS Range (years) 6-69 6-72 No. of males 203 (49.9) 266 (55) NS No. of females 204 (50.1) 214 (45) NS Educational level: Illiterate 24 (6) 41 (9) NS Primary 243 (60) 259 (54) NS Secondary 111 (27) 151 (31) NS Beyond high school 29 (7) 29 (6) NS Occupation: Shepherd 74 (18) 69 (14) NS Student 170 (42) 201 (42) NS Housewife 80 (20) 87 (18) NS Other 83 (20) 123 (26) NS House characteristics: Wood/cemented floor 311 (76) 329 (69) <0.01 1.49 1.09-2.03 Dirt floor 96 (24) 147 (31) <0.02 0.70 0.51-0.95 Adobe walls 224 (55) 298 (62) <0.05 0.75 0.57-0.99 Concrete walls 98 (24) 70 (15) <0.001 1.86 1.30-2.65 Dirt walls 66 (16) 83 (17) NS Rock walls 19 (5) 29 (6) NS Tin roof 386 (95) 445 (93) NS Other 21 (5) 35 (7) NS No. exposed to dogs 245 (60) 331 (69) <0.01 0.68 0.51-0.91 No. exposed to sheep 126 (31) 173 (36) NS NS No. exposed to cattle 195 (48) 234 (49) NS NS a Enzyme-linked immunoelectrotransfer blot assay, portable ultrasonography, and portable chest X-ray. b The non-examination cohort included subjects surveyed using only one or two methods. c Significance by the X2 test. d Figures in parentheses are percentages. e NS: not significant. Twenty-four children under 6 years of age were examined by ultrasound and chest X-ray at the par- ent's request. In a 5-year-old girl who was not part of the EC, a large hydatid cyst in the left lung was found, which was confirmed as hydatid after surgical excision. Two individuals positive by EITB and double-diffusion assays in the 1989 survey had sero- converted to negative and had no demonstrable dis- ease by chest X-ray (in one) and ultrasound (in both). Surgical incidence of hydatidosis. During a 1-year period, only one case of symptomatic pulmonary hydatid disease with haemoptysis presented to the hospital. This individual had previously refused to participate in the survey. A lung hydatid cyst was removed surgically at Oroya Hospital. The symp- tomatic incidence of hydatid disease in Tupac Amaru is 127 per 100 000, which is in sharp contrast to the asymptomatic rate of 2941 per 100 000 de- tected by ultrasound and chest X-ray. Veterinary study The prevalences of hydatid cysts in slaughtered sheep, cattle and pigs are shown in Table 5. Slaugh- terhouse facilities had no veterinary supervision and, WHO Bulletin OMS. Vol 75 1997556 Epidemiology of Echinococcus granulosus infection in Peru Table 2: Age-specific prevalences of hydatidosis in 407 subjects surveyed by enzyme-linked immuno- electrotransfer blot assay (EITB), portable abdominal ultrasound (US), and portable chest X-ray (CXR) in Tupac Amaru, Peru, 1994-95 Screened by: Serology Ultrasound Radiology Positive to EITB Age group No. of (EITB) only: (US) only: (CXR) only: plus US and/or Total (years) subjects EITB + ve Cyst + ve Cyst + ve CXRa positive 6-10b 85 6 (7.Q)c 0 (0) 2 (2.4) 1 (1.0) 9 (10.6) 11-15 74 1 (1.4) 1 (1.4) 0 (0) 3 (4.0) 5 (6.8) 16-30 71 3 (4.2) 1d(1.4) 0 (0) 2 (2.8) 6 (8.5) 31-40 82 1 (1.2) 2 (2.4) 0 (0) 5 (6.1) 8 (9.8) 41-50 54 1 (1.9) 1 (1.9) 1 (1.9) 1 (1.9) 4 (7.4) >50 41 2 (4.9) 2 (4.9) 1 (2.4) 0 (0) 5 (12.2) Total 407 14 (3.4) 7 (1.7) 4 (1.0) 12 (3.0) 37 (9.1) a Three subjects were EITB positive and had lung cysts; eight subjects were EITB positive and had liver cysts, and one was EITB positive and had liver and lung cysts. No patient was EITB negative with liver and lung cysts. b There were significantly more seropositives by EITB in the 6-10-year-old group (X2-test; P < 0.05). c Figures in parentheses are percentages. d One subject had a right pleural effusion detected by both chest X-ray and ultrasound, and two small thoracic cysts were also found by ultrasound. Table 3: Comparison of results of combinations of enzyme-linked immunoelectrotransfer blot assay (EITB), portable abdominal ultrasound (US) examin- ation, and portable chest radiography (CXR) for hydatidosis in 407 subjects, Tupac Amaru, Peru, 1994-95 Results of tests for No. of hydatidosisa subjects EITB/US: +I+ 9b (2.2)c -/+ 7 (1.7) +/- 17d (4.2) 374 (91.9) (Kappa analysis: P = 0.39934; fair agreement range, 0.41-0.60) EITB/CXR: +/+ 4e (1.0) -/+ 4 (1.0) +I- 22 (5.4) -I- 377 (92.6) (Kappa analysis: P = 0.21159; fair agreement range, 0.21-0.40) EITB/US and/or CXR: +/+ 12 (3.0) -/+ 11 (2.7) 14 (3.4) 370 (90.9) (Kappa analysis: P = 0.45725; moderate agreement range, 0.41-0.60) a +, positive for hydatidosis; -, negative for hydatidosis. b One subject with a liver and a lung cyst is included. c Figures in parentheses are percentages. d Three subjects with lung cysts and EITB-positive are included. e One subject with a liver and a lung cyst and three subjects with lung cysts are included. in all our 10 visits, dogs were observed in and around the building feeding on discarded hydatid-infected viscera. A total of 104 dogs (63% of all dogs) were successfully purged; 33 (32%) were found to have adult echinococcus worms. Case-control study A total of 27 cases and 51 controls participated in the case-control study. Ten cases and 23 controls either left the cooperative or declined to participate in the interview. Univariate analysis revealed no significant dif- ferences between cases and controls for the fol- 'lowing variables: level of education, occupation (shepherd or other), time of exposure to dogs, slaughter practices at home, disposal of infected vis- cera, place where animals were slaughtered, and number of dogs kept in the past. Knowledge of how hydatid disease occurred in humans was not a significant variable. Although not different between cases and controls, nearly half of the subjects (46%) had complete or partial knowl- edge about the cause of hydatidosis. A borderline or slight statistical significance was found for the fol- lowing variables: raising sheep in the past, number of dogs owned in the past, owner's leftovers given as food to dogs, use of dogs for shepherding or as pet animals, and lack of knowledge of how hydatidosis was transmitted to humans. WHO Bulletin OMS. Vol 75 1997 557 P.L. Moro et al. Table 4: Presumed hydatid cysts detected by imaging (ultrasound and X-rays) in 23 out of 407 (5.7%) individuals surveyed in Tupac Amaru, Peru, 1994-95 Observations EITBa results Ultrasoundb - Type 3, single hydatid liver cyst in right lobe (5)d + Type 3, single hydatid liver cyst in right lobe (10) + Type 3, multiple cysts in left liver lobe (7), 7 peritoneal cysts (3) Type 1, single hydatid liver cyst in left lobe (12) + Type 1, single liver cyst in right lobe (1.4) + Type 1, single liver cyst in right lobe (13) + Type 3, 5 liver cysts of various sizes, right and left lobes (7,7,4,4,1) + Type 1, single liver cyst in right lobe (15) Type 3, single liver cyst in left lobe (6.3) + Type 1, single liver cyst in right lobe (5) Type 1, single liver cyst in right liver lobe (4) + Type 1, 2 liver cysts in right lobe Type 4, small solid echogenic area in right liver lobe (0.9) + Type 4, small solid echogenic area in right liver lobe (2) - Type 5, calcification in liver (0.5) - Right pleural effusion with two lung cyst images (3) - N.C. - N.C. + N.C. + N.C. N.C. + N.C. - N.C. Radiological No cyst (N.C.) or symptoms N.C. N.C. N.C. N.C. N.C. N.C. N.C. N.C. N.C. N.C. Presumptive hydatid cyst of right lung (3) N.C. N.C. N.C. Right pleural effusion, N.C. Hydatid cyst of right lung (2.5) Hydatid cyst of left lung (6) Presumptive hydatid cyst of left lung (1.3) Presumptive hydatid cyst of right lung (3) Presumptive hydatid cyst of left lung (3.5) Presumptive hydatid cyst of right lung (3.5) Presumptive hydatid cyst of left lung (11) a EITB, enzyme-linked immunoelectrotransfer blot assay; positive (+), negative (-). All EITB-positive results were positive to the three diagnostic bands (M,= 8 x 103, 16 x 103, 21 x 103). b Type 1, collection of fluid without daughter cysts (solitary); type 2, collection of fluid with a split wall; type 3, collection of fluid with septa (daughter cysts); type 4, heterogeneous echographic pattern; type 5, reflecting thick wall (calcified cyst). c Surgically confirmed cases. d Figures in parentheses are the sizes of the cysts in cm. Matched conditional logistic regression analysis was used to assess the independence and strength of variables to identify those individuals at a higher risk of hydatidosis. No statistically significant variables were found. However, those individuals who owned sheepdogs tended to have a higher risk for hyda- Table 5: Prevalence of hydatidosis according to infected organs in sheep, cattle and pigs slaughtered at Tupac Amaru, Peru, 1994 No. of animals examined No. of infected livers No. of infected lungs No. of infected animals Sheep 117 75 (64)a 101 (86) 102 (87) Cattle 25 6 (24) 16 (64) 17 (68) Pigs 8b 70 (88) 70 (88) 7 (88) a Figures in parentheses are percentages. b One animal was infected with cysticerci. c Two animals were heavily infected with cysts in the liver, lung, heart and kidney. WHO Bulletin OMS. Vol 75 1997 Case, sex/age (years) 1,c F/46 2,0 M/45 3,c M/17 4,c F/35 5, M/9 6, F/40 7, M/39 8, F/36 9, M/13 10, M/11 11, F/52 12, M/12 13, M/53 14, M/26 15, M/39 16, M/25 17,c F/43 18,c M/9 19, M/38 20, F/40 21, M/7 22, M/12 23, M/60 Clinical No signs (N.S.) Hepatomegaly N.S. Hepatomegaly N.S. N.S. Hepatomegaly N.S. N.S. N.S. N.S. N.S. N.S. N.S. N.S. N.S. N.S. N.S. N.S. N.S. N.S. N.S. N.S. 558 Epidemiology of Echinococcus granulosus infection in Peru tidosis, and those who did not know how hydatid was transmitted to humans were twice as likely to have hydatidosis. Discussion The 9.1% prevalence of human hydatid infection (as defined, see above) found in this study is the highest reported in Latin America among studies using simi- lar survey techniques (13-16). The 5.7% prevalence of human hydatid disease (again, as defined above) in the Tupac Amaru cooperative in the highlands of Peru is similar to that found in north-western Ken- yan nomadic populations. The Kenyan group, sur- veyed only by portable ultrasound, had the highest human prevalence previously known, 5.6% (17). Additionally, we found an extremely high preva- lence in dogs and livestock which will support the parasite's continued life-cycle and human infection. The serological prevalence of human hydatid disease has more than tripled in the last 5 years (1.9% to 6.4%), which may partly be explained by a more representative population sample and im- provements in the assay. However, it is probable that human prevalence is truly rising. Notably, the preva- lence of infected dogs (the source of human infec- tion) has risen 20-fold since 1980 when the control programme ceased (3, 5). The routine inappropriate disposal of infected viscera is the source of the rising dog prevalence. Presumably this increase in infected dogs creates a more heavily infected environment, which leads to a higher human incidence. The 18- fold rise in the prevalence of infected sheep (3) dur- ing the same period confirms the fact that the environment has become more infectious. The use of ultrasound and chest X-rays in this study has also augmented hydatidosis detection. The majority of cysts detected by these techniques met criteria for a hydatid cyst (8, 9) or were accompanied by positive serology. Therefore, the false-positive rate is likely to be very low. Some limitations of hydatid serology have been confirmed in this study. Assuming that all cysts de- tected by chest X-ray and ultrasound are hydatid, the sensitivity is particularly low in lung cysts (38%) compared to hepatic cysts (57%). This finding is con- sistent with previous observations (7,16-19). The 14 seropositive individuals with negative ultrasound and chest X-ray could be explained by aborted in- fection, undetectably small cysts, or false-positive reactions. Cross-reaction with antibodies to Taenia solium was ruled out by testing these 14 sera with T. solium EITB, all of which were negative (Moro et al., personal observations). Also, given the extremely high specificity of the assay in prior studies, false positives are unlikely (7). Whether these 14 sero- positive, yet radiographically and ultrasonically negative subjects will develop clinically evident dis- ease remains to be seen. Children under 11 years had significantly higher seropositive rates than older subjects, which suggests that transmission is ongoing, frequent, and possibly more intense than occurred previously. This fre- quent exposure to E. granulosus may induce some degree of protection which prevents further develop- ment to the metacestode stage. Early acquired im- a munity may result in aborted infections, and the possibility that this occurs is supported by the finding on ultrasound of atypical images in the liver in three subjects, one of whom was seropositive. In addition, two individuals positive by EITB and double- diffusion assays in the 1989 survey had serocon- verted to negative and had no demonstrable disease by chest X-ray (in one) and ultrasound (in both). An important finding of this study is the high frequency (2%) of lung hydatid cysts in the popula- tion, which is consistent with the high incidence of pulmonary hydatidosis seen in nearby hospitals in Huancayo city and in Lima (20-22). The ratio of lung to liver cysts was 1:1.75. This contrasts with findings in Argentina, Chile, and Uruguay, where the ratio of lung to liver cysts in autopsy series varied from 1:3 to 1:13 (23, 24). A previous chest radiograph survey close to Tupac Amaru revealed a 3% prevalence of presumptive hydatid lung cysts (4). The high fre- quency of lung cysts seen in Peru may be due to the effect of high altitude, which causes an increased volume of capillary blood in the lungs and increased dilation of capillaries (25). The possibility of a local strain of E. granulosus with pulmonary trophism should also be explored. Possible reasons for our inability to demon- strate significant risk factors are that infection may spread beyond a single household and that the study population is relatively homogeneous. Wide geo- graphical dispersal of infection may arise from a single point source of E. granulosus-infected dog faeces, as demonstrated in New Zealand (26). The lapse in hydatid control measures since 1980 is responsible for the higher canine, ovine and human disease prevalence. This increase in echino- coccosis prevalence was thus the result of premature cessation of the control programme. Acknowledgements This study was partly funded by NIH grant No. U01A135894-01 and by CONCYTEC, Lima, Peru, which was financed in part by PAHO. We thank J. Moro, N. Perez-Palma, J.B. Phu, D. Sara, P.M. Schantz and L. Moulton for editorial help, and WHO Bulletin OMS. Vol 75 1997 559 P.L. Moro et al. C. Fillmore for aid in data analysis; Dr M. Gamarra, R. Zarate, and the population of Tupac Amaru for their coop- eration; and Dr G. Leguia, Dr San Martin (IVITA, UNMSM), Dr A. Acosta (PSO-IPSS-Huancayo), Dr 0. Nieto (Pachacayo-IPSS), V. Gutierrez, L. Verastegui, M. Calderon, and the following members of the Cysticercosis Working Group in Peru (CWG): H.H. Garcia, M. Tovar, V.C.W. Tsang, P. Torres, G. Madico, A. Gonzalez, C. Gavidia and L. Cabrera. Resume Epid6miologie de l'infection A Echinococcus granulosus dans la partie centrale des Andes p6ruviennes L'hydatidose humaine due a Echinococcus granulo- sus constitue un grave probleme de sante publique dans les regions du Perou ou se pratique l'elevage des ovins. Dans les annees 70, un programme pilote de lutte contre I'hydatidose comportant une purgation des chiens par le bromhydrate d'areco- line suivie d'un traitement par le praziquantel, a e conduit dans la cooperative agricole Tupac Amaru, dans la partie centrale des Andes peruviennes. Les membres de la cooperative avaient en outre 6te invites a ne pas nourrir les chiens avec des abats crus. Ces mesures ont toutefois ete abandonnees en 1980 et il n'existe pas de donnees completes sur I'ampleur de l'infection par le tenia echinocoque chez les h6tes d6finitifs, interm6diaires et accidentels. La presente etude avait pour but de determiner la pre- valence de l'echinococcose humaine et animale a E. granulosus et d'identifier les facteurs environne- mentaux et culturels jouant un role important dans la transmission de l'infection. La pr6valence de l'6chinococcose humaine a ete determinee par echographie et radiographie pulmonaire au moyen d'appareils portables, et par serologie au moyen d'une methode immuno- enzymatique (EITB). La pr6valence de l'infection animale a ete determinee chez le chien par examen microscopique des selles pour y rechercher les tenias apres purgation par l'arecoline, et chez les ovins par examen des visceres apres abattage. Les sujets trouv6s positifs a l'un quelconque des examens ci-dessus (cas) et des sujets non infect6s provenant de la meme communaut6 (temoins) ont ete inclus dans une etude cas-temoins destin6e a identifier les facteurs importants dans la transmis- sion de l'infection a E. granulosus. La prevalence de l'infection (sujets positifs pour au moins un des examens mentionnes ci- dessus) parmi les 407 personnes incluses dans l'etude 6tait de 9,1% (37/407). La pr6valence de I'hydatidose humaine (r6sultat positif a l'6cho- graphie et/ou la radiographie pulmonaire) etait de 5,7% (23/407). La localisation etait hepatique dans 3,4% des cas, pulmonaire dans 2,0% des cas et A la fois hepatique et pulmonaire dans 0,2% des cas. Le rapport des localisations pulmonaires/ h6patiques etait de 1:1,75. Les enfants de moins de 11 ans avaient des taux de seropositivit6 signi- ficativement plus eleves que les adultes, sans signes de maladie 'a 'examen echographique ou radiographique pratique au moyen d'un appareil portable (p < 0,05). L'analyse par regression logis- tique conditionnelle n'a pas fait apparaitre de varia- bles statistiquement significatives. Cependant, les personnes qui possedaient des chiens de berger tendaient a avoir un risque plus eleve d'hydatidose, et celles qui ne savaient pas comment la maladie se transmet 'a 'homme avaient un risque multiplie par deux. Parmi les 104 chiens examines apres purga- tion par l'arecoline, 33 (32%) hebergeaient des tenias adultes. Parmi les 117 moutons abattus a I'abattoir local, 102 (87%) etaient porteurs de kystes hydatiques. La pr6valence de l'hydatidose humaine dans cette zone d'endemie est l'une des plus elevees au monde et est cinq fois plus elevee que la valeur rapportee auparavant dans cette meme zone. La prevalence de 5,7% de I'hydatidose relevee sur les hauts plateaux du Perou dans la cooperative agricole Tupac Amaru est similaire a celle que l'on observe chez les populations nomades du nord- ouest du Kenya. Nous avons 6galement observ6 une forte prevalence de l'infection chez les chiens et les ovins, avec des taux multiplies respec- tivement par 20 et par 18 depuis l'arret des mesures de lutte en 1980. Les enfants de moins de 1 1 ans ont des taux de seropositivite significativement plus eleves que les autres groupes d'ages, ce qui laisse a penser que la transmission se poursuit actuel- lement avec une grande frequence et peut-etre une plus forte intensite qu'auparavant. L'absence de mesures de lutte depuis 1980 est responsable de I'augmentation de la prevalence de I'hydatidose chez le chien, les ovins et I'homme. Cette situation est due a l'arret pr6matur6 des pro- grammes de lutte. References 1. Otarola SG. [Epidemiology of hydatidosis in Peru]. Boletin de la Oficina Sanitaria Panamericana, 1966, 60:144-153 (in Spanish). 2. Zapatel J, Guerrero C, Escalante J. [Hydatidosis in Peru]. Boletin de la Oficina Sanitaria Panamericana, 1962, 52: 296-308 (in Spanish). 560 WHO Bulletin OMS. Vol 75 1997 Epidemiology of Echinococcus granulosus Infection in Peru 3. Ministry of Health. [National Zoonosis' Control Programme. Records of the National Seminar on Zoonoses and Foodborne Diseases]. Ministerio de Salud, 3-4 July 1989, Lima, Peru: 122-137 (in Spanish). 4. Varela Diaz VM et al. [Hydatidosis in Peru: impor- tance of immunodiagnosis and of seroepidemiological studies]. Acta m6dica peruana, 1984, 11: 21-28 (in Spanish). 5. Arevalo F et al. [Prevalence and control of echino- coccosis in a pilot programme in the Sierra Central of Peru]. Paper presented at the VII National Congress of Veterinary Sciences, Ica, Peru, November 1982 (in Spanish). 6. Moro PL et al. Distribution of hydatidosis and cysticercosis in different Peruvian populations as demonstrated by an enzyme-linked immunoelectro- transfer blot (EITB) assay. American journal of tropical medicine and hygiene, 1994, 51: 851-855. 7. Verastegul M et al. Enzyme-linked immunoelectro- transfer blot test for the diagnosis of human hydatid disease. Joumal of clinical microbiology, 1992, 30: 1557-1 561. 8. Gharbi HA et al. Ultrasound examination of hydatic liver. Radiology, 1981, 139: 459-463. 9. Pant CS, Gupta RK. Diagnostic value of ultrasono- graphy in hydatid disease in abdomen and chest. Acta radiologica, 1987, 28: 743-745. 10. Bernstein J. A classification of renal cysts. In: Gardner KD, ed. Cystic diseases of the kidney. New York, John Wiley, 1976: 7. 11. Beggs I. The radiology of hydatid disease. American joumal of roentgenology, 1985, 145: 639-648. 12. FAO/UNEP/WHO. Parasitological diagnosis of adult tapeworm in carnivores. In: Eckert J, Gemmell MA, Soulsby EJL, eds. Guidelines for surveil- lance, prevention and control of echinococcosisl hydatidosis. Geneva, World Health Organization, 1981: 70-78. 13. Frider B et al. [Echographic, serologic and radiologic registry of human hydatidosis: contribution to a control programme]. Acta gastroenterologica latinoameri- cana, 1985, 15:199-211 (in Spanish). 14. Perdomo R et al. Early diagnosis of hydatidosis by ultrasonography, 1988, Lancet, 1: 244. 15. Larrieu E et al. [Human hydatidosis: field echography for the determination of groups at high risk in the evaluation of a control programme]. Revista del Instituto de Medicina Tropical de Sao Paulo, 1989, 31: 267-270 (in Spanish). 16. Barbieri M et al. Use of specific antibody and circu- lating antigen serum levels in the hydatid immuno- diagnosis of asymptomatic population. International journal for parasitology, 1994, 24: 937-942. 17. Macpherson CNL et al. Portable ultrasound scanner versus serology in screening for hydatid cysts in a nomadic population. Lancet, 1987, 2: 259-291. 18. Varela Diaz VM, Guarnera EA, Coltorti EA. [Advan- tages and limitations of the immunological and image- detection methods for the diagnosis of hydatidosis]. Boletin de la Oficina Sanitaria Panamericana, 1986, 100: 369-383 (in Spanish). 19. Craig PS, Zeyhle E, Romig T. Hydatid disease: re- search and control in Turkana. II. The role of immu- nological techniques for the diagnosis of hydatid disease. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1986, 80: 183-192. 20. Chahud IA, Wurs G. [Hydatid disease of the liver: epidemiological, clinical and therapeutic aspects]. Revista gastroenterol6gica del Peru, 1985, 5: 113- 127 (in Spanish). 21. Rodriguez T. [Epidemiological study of human hydatidosis reported during the period 1975-1986 in the principal hospitals of metropolitan Lima]. Un- published thesis. Faculty of Veterinary Medicine, Universidad Nacional Mayor de San Marcos, Lima, Peru, 1990 (in Spanish). 22. Alarcon J et al. [Pulmonary hydatidosis: epidemio- logical study of urban cases in the Hipolito Unanue Hospital of Lima]. Revista peruana de epidemiologia, 1992, 5:15-19 (in Spanish). 23. Purriel P, Mendoza G, Decedo H. [Hydatidosis in Uruguay: epidemiological study (1962-1968)]. Torax, 1990,19: 1-15 (in Spanish). 24. Schenone H, Rojas A, Ramirez R. [Frequency of hydatidosis in autopsies carried out in the Instituto de Medicina Legal and Hospital of Santiago, Chile (1947-1970)]. Boletin chileno de parasitologia, 1971, 26: 98-103 (in Spanish). 25. Lahiri S. Physiological responses and adaptations to high altitude. International review of physiology, 1977, 15: 217-251. 26. Gemmell MA, Lawson JR. Epidemiology and control of hydatid disease. In: Thompson RCA, ed. The biol- ogy of echinococcus and hydatid disease. London, George Allen & Unwin, 1986: 189-211. WHO Bulletin OMS. Vol 75 1997 561