Prevalence of Endoparasitic Infection in Children and Its Relation with Cholera Prevention Efforts in Mexico

Prevalence of endoparasitic infection in children and its relation with cholera prevention efforts in Mexico

Charles T. Faulkner,1 Benito Borrego Garcia,2 Michael H. Logan,3 John C. New,1 and Sharon Patton1

ABSTRACT Objective. To investigate whether increased knowledge and use of public health measures promoted for cholera prevention is reflected in lower prevalence of parasitic infection in households in a community in the state of Tamaulipas, Mexico, that is close to the border with the United States of America. Methods. Between 1994 and 1997, fecal samples from 438 children were collected through convenience sampling and then examined for helminth eggs/larvae and protozoan cysts as bio- logic indicators of household compliance with recommended cholera prevention measures. The suggested measures were to wash hands before meals and after defecation, to drink purified water, to wash fruits and vegetables, and to eat well-cooked food. In addition, information on the knowledge of and the use of cholera preventive measures was collected by interviews with adult informants in 252 households (186 of those households also provided a fecal sample for analysis). Results. Parasitic infections occurred in 131 of the 438 children (30%), who resided in 79 of the 186 households (42%) that provided fecal samples. Giardia lamblia accounted for 12.5% of all infections. Infections with Hymenolepis nana, Ascaris lumbricoides, Trichuris trichiura, Enterobius vermicularis, Ancylostoma/Necator, Strongyloides stercoralis, En- tamoeba coli, Entamoeba hartmanni, Entamoeba histolytica, Endolimax nana, and Io- damoeba bütschlii were also noted. Infected children were older and more often had an infected sibling. Households with three or more children were also more likely to have an infected child. The primary caregivers in the households where at least one child had a parasitic infection were distinguished by their inability to list at least three cholera prevention measures from memory. Conclusions. The 42% household prevalence of parasitic infection was relatively high and indicates that some residents of this community may not have fully embraced the public health education efforts promoted for prevention of cholera. The occurrence of nonpathogenic protozoan parasites such as Endolimax nana, Entamoeba coli, Entamoeba hartmanni, and I. bütschlii are important bioindicators for the persistence of unhygienic behaviors that increase the risk of cholera and other infectious diseases dependent on fecal-oral transmission. Informa- tion obtained by similar studies can be useful for monitoring compliance with community health and hygiene programs and may indicate the need to intensify educational efforts for the prevention of diarrhea associated with enteric pathogens that cannot be controlled by drugs alone.

Key words Cholera; parasitic diseases; health education; knowledge, attitudes, practice; Mexico.

1 University of Tennessee, Department of Compara- University of Tennessee, 2407 River Drive, Knox- 2 Médico veterinario zootecnista, Valle Hermoso, tive Medicine, Knoxville, Tennessee, United States ville, Tennessee 37796-4543, United States of Amer- Tamaulipas, Mexico. of America. Send correspondence to: Charles T. ica; telephone: (865) 974-5718; fax: (865) 974-5640; 3 University of Tennessee, Department of Anthropol- Faulkner, Department of Comparative Medicine, e-mail: [email protected] ogy, Knoxville, Tennessee, United States of America.

Rev Panam Salud Publica/Pan Am J Public Health 14(1), 2003 31 The epidemic spread of Vibrio cho- community would be reflected in lower rounding communities and establish lerae, biotype El Tor, serotype Inaba, occurrence of infection with intestinal homes in neighborhoods developing into Latin America in 1991 prompted a protozoan and helminth parasites be- on the outskirts of Valle Hermoso. Con- vigorous response from public health cause their transmission is facilitated structed of concrete block or wood agencies at the international, national, by the same unhygienic behaviors that frame with plywood or flat-board sid- state, and local community levels (1–4). increase risk for cholera. ing, the houses in these neighborhoods Education materials disseminated These unhygienic behaviors include have between one and three rooms. to the public described the nature of inadequate hand-washing, drinking Electrical utilities and running water in- the disease and provided recommen- contaminated water, eating improp- side the house are available in approxi- dations for prevention of infection by erly washed fruits and vegetables, and mately 90% of the households (7). Out- cholera and similar pathogens. Many promiscuous defecation habits. The oc- side latrines are the primary method for residents of these countries became currence of nonpathogenic protozoan disposal of fecal waste, and only 19% of acutely aware of the fecal-oral route of parasites such as Endolimax nana, Io- the households have indoor toilets (7). cholera transmission and the impor- damoeba bütschlii, Entamoeba coli, and The decision to conduct this inves- tance of personal and household hy- Entamoeba hartmanni in populations tigation in Valle Hermoso was moti- giene measures in maintaining health. targeted by interventions to prevent vated primarily by our preexisting Increased promotion of personal hy- cholera are important bioindicators for relationship with local government giene measures such as frequent hand- unhygienic behavior because these officials and health department per- washing, consumption of bottled or parasites do not produce clinical dis- sonnel. The relatively small size of the chemically treated water, and thor- ease and are not treated by physicians. community and its convenient dis- oughly washing fruits and vegetables Use of these parasites as bioindicators tance from the border with the United prior to consumption helped mitigate of unhygienic behavior provided us States were secondary considerations the adverse impact of the cholera epi- with a method for objectively evaluat- that made this an attractive place for demic throughout Latin America, in- ing the quality of information on the the investigation. cluding Mexico (3–4). These efforts may knowledge and practice of cholera pre- have also had an additional effect on ventive measures obtained from inter- the community—wide prevalence of views with the primary caregivers in Household enlistment infection with endoparasitic helminths participating households. This investi- and data collection and protozoa dependent on the fecal- gation offered a unique opportunity to oral transmission route. De la Sotta examine an important dimension of All of the households in the study and colleagues (5) suggested that the human behavior by comparing what were located in neighborhoods that Chilean cholera control program was informants say they do with what they had been the focus of recent public responsible for a statistically signifi- actually do. The information obtained health education initiatives to reduce cant decrease in intestinal parasitoses through this research can be used by childhood morbidity from cholera and diagnosed at hospitals in the city of public health workers to monitor com- diarrheal disease produced by para- Santiago during the 3-year period that pliance with community health and sitic and other infectious enteric agents. followed a national campaign. How- hygiene programs, and it may indicate The ongoing educational effort was ever, this association may have been the need to intensify educational ef- initiated in 1993, that is, the year be- coincidental because Santiago resi- forts for prevention of diarrhea associ- fore the start of this investigation, and dents were not interviewed to deter- ated with enteric pathogens such as it was carried out by local committees mine if they knew or used any of the cholera, Salmonella, Cryptosporidium sp., of volunteer health promoters trained anti-cholera recommendations. and Giardia lamblia that cannot be con- to work under the supervision of We undertook our research to inves- trolled by drugs alone. nurses from a Valle Hermoso health tigate the impact that a locally admin- center operated by the Government’s istered program to prevent cholera had Secretariat of Health (Secretaría de Sa- on the prevalence of endoparasitic in- METHODS lud). Having the health promoters fection in households of a small com- structure already in place maximized munity in the state of Tamaulipas, in This study was carried out between our access to households and helped northeastern Mexico. Like de la Sotta 1994 and 1997 in Valle Hermoso, Ta- assure rapport with subjects. and colleagues (5), we were intrigued maulipas, Mexico, a community of Some of the participants enlisted in by the possibility that increased com- 39 004 persons located 45 km south the study during open meetings with munity awareness and utilization of of the border with the United States of health educators and two of the au- cholera prevention measures could re- America (6–7). Employment and in- thors (CTF and BBG). At these meet- duce the prevalence of intestinal para- vestment opportunities arising from ings the project objectives were de- sitic infections. We hypothesized that recent industrial growth and increased scribed by the two authors (CTF and household compliance with the edu- trade with the United States have BBG), and the participants received in- cational initiative undertaken in this prompted people to relocate from sur- structions for collecting fecal speci-

32 Faulkner et al. • Prevalence of endoparasitic infection in children and cholera prevention efforts in Mexico mens from resident children for para- Use of the preventive measures in infected sibling. The occurrence of en- sitologic analysis. Other participants the households was classified as con- doparasitic infection in households were recruited during neighborhood sistent based on the response that they was compared by the number of resi- public health clinics conducted un- were used all of the time, or inconsis- dent children, ability of the PC to recall der the auspices of the Valle Hermoso tent based on the response that they cholera prevention measures, self- health center. were either used some of the time or reported frequency of their use, drink- Participation in the study was strictly never used. ing water source, and household facili- voluntary, with no monetary incen- When the questionnaire was admin- ties for fecal waste disposal. Participants tives offered. Informed consent for all istered, each PC received explicit in- consisted of children ≤ 16 years old household participants was obtained structions on how to collect fecal sam- (n = 438) from whom fecal samples in compliance with the University of ples from their children and to label, were examined for parasitic infection Tennessee Institutional Review Board for each sample submitted, plastic by collection on a single occasion, and guidelines for research involving hu- bags (provided by the authors) with the household PCs (n = 242) inter- man subjects. Participants were in- the child’s name, sex, age, family viewed with the household health and formed that all children who tested name, and name of the neighborhood. hygiene questionnaire administered by positive for clinically significant par- The fecal samples were collected by members of the research team. asitic infections would be referred to the authors 24 hours later, taken to a Data were analyzed by appropriate the Valle Hermoso health center, where field laboratory, divided into equal parametric and nonparametric statisti- treatment would be provided at no cost. portions and transferred to fecal trans- cal methods with the InStat 3 software The households surveyed in this study port vials (Meridian Diagnostics, Cin- package (GraphPad Software, San constituted a convenience sample. cinnati, Ohio, United States) that con- Diego, California, United States) for Small teams of neighborhood health tained either 10% buffered neutral the Microsoft Windows 95 operating promoters, nurses, and the authors formalin (BNF) or polyvinyl alcohol system. Unpaired Student’s t tests interviewed household informants (PVA), and then transported to the were used to test for any significant during door-to-door surveys. Ques- University of Tennessee Clinical Par- difference in the mean age of infected tionnaires were used to interview the asitology Laboratory. Samples pre- and uninfected children, and for age- primary caregiver (PC) in each partici- served in BNF were analyzed for diag- related differences in infection with pating household. The information nostic stages of endoparasites by the specific parasite species. Single factor that was collected included: age and centrifugal flotation method, with sat- analysis of variance (ANOVA) was sex of the informant; the age, the sex, urated zinc sulfate (specific gravity used to test for differences in mean age and the number of resident children; 1.18) and Sheather’s sucrose (specific among informants from three groups: source of drinking water; and facilities gravity 1.26) solutions as flotation infected households (n = 79), uninfected for disposal of fecal waste. Informa- media (8). Eggs and cysts of helminths households (n = 107), and a smaller tion was also collected on household and protozoa were identified based group of informants whose household income, based on the occupation of the on their characteristic morphology infection status was unknown because head of the household, and on home- and size, with the aid of diagnostic ref- they declined to submit fecal samples ownership. Knowledge of cholera pre- erence manuals (8–11). Fecal smears for parasite analysis (n = 56). ANOVA vention measures was assessed by made from PVA-fixed samples were was also used to test for any difference asking informants to recall these rec- stained with the modified trichrome in the weekly mean income of these ommendations from memory: staining technique and examined to three household groups. The statistical confirm the identification of protozoan association of parasitic infection with 1) wash hands after defecation and cysts recognized from the zinc sulfate other nominal and discrete variables before meals flotation (11). All samples were also (e.g., residing in the household with 2) wash fruits and vegetables before macroscopically examined for blood, an infected sibling, the number of res- consumption mucus, spontaneously passed worms, ident children, and the ability of the 3) cook all foods well, especially tapeworm segments, and fly larvae. PC to recall cholera prevention mea- seafood sures) was evaluated with Fisher’s 4) only drink water that has been exact test. Level of significance (α) was boiled or purified by filtration or Study design, analytical units, set a priori at ≤ 0.05. Prevalence odds chemical treatment and statistical methods ratios were calculated to estimate the strength of association between the oc- Each informant was asked, “Do you This investigation was a cross-sec- currence of parasitic infection and sus- know the government recommenda- tional study based on convenience pected exposures (e.g., the presence of tions for the prevention of cholera?” sampling (12). Endoparasitic infections an infected child in the household, in- Informants answering in the affirma- in children (i.e., positive or negative) ability to recall at least three cholera tive were then asked to list the recom- were compared by age and sex of the prevention measures, or inconsistent mendations from memory. child and the presence or absence of an use of preventive measures) (12).

Rev Panam Salud Publica/Pan Am J Public Health 14(1), 2003 33 RESULTS were children without this parasite Parasitic infection in households (Table 2). Parasites found in children Parasitic infections in children older than the overall mean age were The occurrence of parasitic infection Hymenolepis nana, Trichuris trichiura, in households was based on 186 house- Fecal samples of 438 children were Enterobius vermicularis, Entamoeba coli, holds that provided fecal samples for analyzed for endoparasitic infection Endolimax nana, and Iodamoeba bütschlii analysis. Of the 186 households, 79 of (Table 1). The mean age of the children (Table 2). However, this association them (42%) had at least one infected was 6 years. The number of fecal sam- was not statistically significant. More child. The other 107 households were ples was approximately equal for males of the children infected with Ascaris considered uninfected based on a sin- and females, and there was no signifi- lumbricoides and Giardia lamblia were gle negative fecal examination from all cant difference in age by sex (Table 1). younger than the overall mean age, children residing in the household Evidence of parasitic infection was but this association was also not statis- (Table 4). Informants from these 186 found in 30% of all examined samples tically significant (Table 2). Infected households provided information on (Table 1). Endoparasitic infections were children were 2.9 times as likely to their knowledge and use of cholera indicated by the occurrence of helminth have an infected sibling as were unin- prevention measures and on other eggs or larvae or protozoan tropho- fected children (Table 3). characteristics relevant to understand- zoites or cysts in a fecal sample col- Infections with two or more para- ing the distribution of parasitic infec- lected on a single occasion. Infection sites occurred in 31% of all positive tions in their children. This information status was not associated with the sex children tested (Table 1). The occur- was also obtained from 56 additional of the child despite the observation rence of multiple infections was not informants whose household infection that slightly more female children had associated with the age or sex of the status was unknown because they de- positive fecal examinations than did child (Table 1) or with the infection sta- clined to submit fecal samples from male children (Table 1). Infected chil- tus of their siblings (Table 3). The sam- their children following the interview. dren were significantly older than were ple sizes did not allow for analysis of Although these informants cannot be uninfected children (Table 1). parasite species distributions within regarded as a “control” group in the In terms of specific parasites, chil- households or neighborhoods or of true sense of the definition, they con- dren infected with Entamoeba hart- their co-occurrence with other parasite tributed an additional dimension for manni were significantly older than species. describing the demographic character-

TABLE 1. Age and gender data on 438 children ≤ 16 years old analyzed for endoparasitic infection, Valle Hermoso, Tamaulipas Mexico, 1994–1997

Probability of statistical Males Females All childrena association

Total fecal samples examined (no., %) 223 (50.9%) 215 (49.1%) 438 (100%) NAb Samples positive for parasitic infection (no., %) 63c (48.1%) 68c (51.9%) 131 (29.9%) 0.46c Positive for infection with single parasite species 44 47 91 NA Positive for infection with multiple parasite species 19d 21d 40 (30.5%) 1.00d Samples negative for parasitic infection (no., %) 160 (52.1%) 147 (47.9%) 307 (71.1%) NA Mean age irrespective of infection status (yr) (SD)e 6.04 (3.29)f 5.86 (3.68)f 5.95 (3.48) 0.59f Mean age of infected children (yr) (SD) 6.47 (3.08) 6.47 (3.78) 6.47 (3.45)g 0.04g Mean age of children with single parasite infections (yr) (SD) 6.35 (2.95) 6.81 (4.17) 6.59 (3.62)h NA Mean age of children with multiple parasite infections (yr) (SD) 6.73 (3.41) 5.72 (2.69) 6.20 (3.05)h 0.56h Mean age uninfected children (yr) (SD) 5.85 (3.37) 5.55 (3.60) 5.71 (3.88)g NA a Calculation of mean age is based on 410 children (129 infected, 281 uninfected) for whom age was recorded. b NA = not applicable (statistical comparison not performed or the associated probability value is placed in another cell with matching superscript to indicate compared values). c The frequency in the number of samples positive for parasitic infection between male and female children was not statistically significant for nonrandom association by Fisher’s exact test (P = 0.46, degrees of freedom (df) = 1). d The frequency in the number of samples positive for infection with multiple parasitic species between male and female children was not statistically significant for nonrandom association by Fisher’s exact test (P = 1.00, df = 1). e SD = standard deviation associated with calculated mean age of children from whom samples were examined. f Differences in mean age of male and female children from whom samples were examined were not statistically significant by two-tailed Student’s t test for unequal groups (P = 0.59, df = 408). g The difference in mean age of children infected with any parasite compared to uninfected children was statistically significant by two-tailed Student’s t test for unequal groups (P = 0.04, df = 408). h The difference in mean age of children infected with a single parasitic species compared to children infected with multiple parasitic species was not statistically significant by two-tailed Stu- dent’s t test for unequal groups (P = 0.56, df = 127).

34 Faulkner et al. • Prevalence of endoparasitic infection in children and cholera prevention efforts in Mexico TABLE 2. Frequency of endoparasitic infections in children ≤ 16 years of age) by age, Valle Hermoso, Tamaulipas, Mexico. 1994–1997

Age (yr) of Age (yr) of Samples infected children uninfected children Probability of inequality in Parasite species Positive/Examined % Meana SD b Meana SD b mean age

Protozoa Giardia lamblia 55/438 12.55 5.70 3.17 5.99 3.53 0.582 Entamoeba coli 27/438 6.16 6.87 3.40 5.89 3.49 0.158 Entamoeba hartmanni 24/438 5.47 8.22 3.89c 5.81 3.41c 0.001c Entamoeba histolytica 1/438 0.22 NAd NA NA NA NA Endolimax nana 23/438 5.25 6.57 3.06 5.91 3.51 0.379 Iodamoeba bütschlii 10/438 2.28 7.70 4.00 5.91 3.47 0.109 Helminths Hymenolepis nana 28/438 6.39 7.07 2.72 5.87 3.52 0.079 Ascaris lumbricoides 16/438 3.65 4.50 2.37 6.01 3.51 0.0 Trichuris trichiura 6/438 1.36 6.83 4.26 5.94 3.48 0.535 Enterobius vermicularis 6/438 1.36 6.16 1.83 5.95 3.50 0.881 Ancylostoma/Necatore 1/438 0.22 NA NA NA NA NA Strongyloides stercoralis 1/438 0.22 NA NA NA NA NA a Calculation of mean age based on 410 children (129 infected, 281 uninfected) for whom age was recorded. b SD = standard deviation. c Statistically significant difference in mean age by two-tailed Student’s t test for unequal groups (P < 0.001, degrees of freedom = 408). d NA = not applicable; mean age of infection with Entamoeba histolytica, Ancylostoma/Necator, and Strongyloides stercoralis was not calculated because of occurrence as individual infections each. e The eggs of these genera are morphologically indistinguishable.

TABLE 3. Frequency of parasitic infection in children and households according to residence in a household with an infected sibling, and relative number of children in household, Valle Hermoso, Tamaulipas, Mexico, 1994–1997

Prevalence 95% confidence limit Parasite infection status Residence in household with: odds ratio Lower Upper

Infected sibling Uninfected sibling Child infected 75 38 2.882a 1.808 4.592 Child uninfected 100 146

Infected sibling Uninfected sibling Children with multiple parasite species 24 49 1.020b 0.483 2.373 Children with single parasite species 12 25

3 or more children 1 or 2 children Household infected 50 29 4.637c 2.481 8.668 Household uninfected 29 78 a The frequency of infected children residing in households with infected siblings compared to uninfected children residing in households with uninfected siblings was statistically significant for nonrandom association by Fisher’s exact test (P < 0.0001, degrees of freedom (df) = 1). b The frequency of children infected with multiple parasitic species and residence in a household with an infected sibling was not statistically significant for nonrandom association by Fisher’s exact test (P = 1.00, df = 1). c. The frequency of infected children residing in households with three or more children compared to uninfected children residing in households with two and fewer children was statistically significant for nonrandom association by Fisher’s exact test (P < 0.0001, df = 1).

istics and range of information pro- did not differ appreciably from the absence was not associated with house- vided by the study participants. participants (n = 56) whose household hold infection status (Table 4). Married females, who had a mean infection status was unknown (Table Infection status was significantly as- age of 31 years, accounted for 90% of 4). Mean weekly household income sociated with the number of children the persons interviewed in the three and homeownership for participants residing in the household. Households groups of household informants (Table in the study did not differ significantly with three or more children were 4.6 4). These demographic characteristics among the three groups (Table 4). Al- times as likely to have an infected were similar for informants from in- though flush toilets were more fre- child as compared to households with fected and uninfected households and quent in uninfected households, their fewer children (Table 3).

Rev Panam Salud Publica/Pan Am J Public Health 14(1), 2003 35 TABLE 4. Characteristics of informants and households studied for parasitic infection and its relation to cholera prevention measures in Valle Hermoso, Tamaulipas, Mexico. 1994–1997

Household infection status Probability of Infected Uninfected Unknowna statistical Informant and household characteristics (n = 79) (n = 107) (n = 56) association

Household informants Mean age (yr) (SD) b 31.40 (8.16) 31.37 (8.58) 31.83 (11.83) 0.951c Female (no., %) 73 (92%) 96 (90%) 50 (89%) NAd Male (no., %) 6 (8%) 11 (10%) 6 (11%) NA Mean weekly incomee (pesos) (SD) 215.26 (79.56) 216.93 (86.22) 235.75 (105.03) 0.352f Homeownership (no., %) 58 (74%) 74 (69%) 39 (69%) 0.60g

95% confidence Prevalence limit Facilities for fecal waste disposal Infectedh Uninfectedh odds ratio Lower Upper

Outside latrine 63 77 1.86h 0.849 4.07 Flush toilet in house 11 25 a Household infection status was unknown because informants from these households declined to submit fecal samples for parasite analysis. b SD = standard deviation. c Difference in mean age of informants from households of infected, uninfected and unknown infection status was not statistically significant by single factor analysis of variance (ANOVA) (P = 0.951, degrees of freedom (df) = 239). d NA = not applicable, statistical comparison not performed. e The exchange rate for pesos was 7.3 pesos = US$ 1.00 as of January 22, 1996. f Difference in mean weekly income of households of infected, uninfected, and unknown infection status was not statistically significant by ANOVA (P = 0.352, df = 239). g Difference in the frequency of homeownership among households of infected, uninfected, and unknown infection status was not significantly different for nonrandom association by chi-square test (P = 0.60, df = 2). h Facilities for fecal waste disposal not reported for all households. Household infection status and access to facilities for fecal waste disposal was not statistically significant for nonrandom association by Fisher’s exact test (P = 0.13, df = 1).

Parasitic infection and calle), and that following advice from not be taken directly from the faucet knowledge and use of cholera nurses, brochures, and radio broad- without treatment or filtration. Al- prevention measures casts was an important intervention though informants from 5 of the 145 for cholera prevention. These preven- households mentioned that the addi- Informants in each of the participat- tive measures were listed by single in- tion of cloro (i.e., household bleach, ing household groups were knowl- formants and could not be consistently 5.25% sodium hypochlorite) was an edgeable regarding the recommenda- categorized with respect to the other important method for treating drink- tions promoted for cholera prevention thematic classes. ing water, only a single informant was as indicated by their ability to provide Recall of these general thematic able to recall from memory that two at least one preventive measure from classes of cholera preventive measures drops of cloro per liter were required to memory. Seventy-five percent of all was similar between informants from adequately disinfect drinking water. persons interviewed were able to re- infected and uninfected households Personal hygiene measures recalled call at least one preventive measure (Table 5). Informants from 88 unin- from memory included hand-washing from memory. The open-ended nature fected households were able to recall a practices before meals and after using of our request to list the cholera pre- total of 241 preventive measures from the bathroom. The importance of us- ventive measures produced many in- memory, and informants from 57 in- ing designated latrines for defecation dividual responses that were themati- fected households listed 131 total pre- and the admonishment that defecation cally consistent with three general ventive measures (Table 5). However, should not be done freely in the envi- classes of interventions related to the frequency with which preventive ronment was also considered as a pre- treatment of drinking water, personal measures from the thematic classes ventive measure associated with per- hygiene (e.g., hand-washing before were listed by informants was not sig- sonal hygiene. meals and following defecation), and nificantly associated with household Food hygiene measures recalled by food hygiene (e.g., washing fruits and infection status (Table 5). informants included the recommenda- vegetables, and cooking all food well). Drinking water treatments recalled tions of washing fruits and vegetables A fourth class of “other preventive from memory by informants included so as to remove contaminants and of measures” included the admonition to boiling, filtering, adding a small cooking all food well. Hand-washing avoid consumption of food offered by amount of chlorine bleach, and the ad- that was mentioned specifically in con- street vendors (no come comida en la monition that drinking water should nection with food preparation (e.g.,

36 Faulkner et al. • Prevalence of endoparasitic infection in children and cholera prevention efforts in Mexico before preparing or cooking food) was inconsistent use of cholera preventive other pathogens associated with diar- considered a preventive measure asso- measures in their households were 3.7 rheal disease (2, 4). Widespread adop- ciated with food hygiene. The basis for times as likely to obtain their drinking tion of recommendations to frequently this distinction was that hand-washing water directly from the faucet without wash hands following defecation and prior to food preparation prevented treatment (Table 7). prior to meal consumption, drink wa- the spread of infection to others, while ter that has been purified or treated, hand-washing before meals or after and eat well-cooked food may have using the bathroom was important for DISCUSSION been influential in the low occurrence prevention of infection to oneself. of parasitic infection observed in the Household infection status was not The 1991 threat of epidemic cholera northeastern Mexican community re- statistically associated with the fre- provided Government public health ported in this study. quency of the different classes of chol- authorities in Mexico with the impetus Promoting personal hygiene mea- era preventative measures recalled to implement a community-based in- sures for preventing cholera infection (Table 5). tervention strategy for promoting per- in Chile was believed to be responsible The ability to list at least three pre- sonal and food hygiene measures to for reducing the prevalence of infec- ventive measures was significantly prevent infection with cholera and tion with intestinal parasites that de- associated with household infection status (Table 6). Informants who were unable to list at least three cholera preventive measures were two times as TABLE 5. Frequency of cholera preventive measures (grouped according to general the- likely to have an infected child in the matic class) recalled from memory by 145 household informants during interviews, Valle household (Table 6). Household infec- Hermoso, Tamaulipas, Mexico, 1994–1997 tion status was not statistically associ- Household parasite ated with informant’s self-reported infection status consistent or inconsistent use of Infected Uninfected cholera prevention measures (Table 6). Thematic class of cholera (n = 57) (n = 88) Probability of However, households with parasitic prevention measure No. % No. % statistical association infections were five times as likely to obtain their drinking water directly Treatment of drinking water 32 24 44 18 from the faucet without treatment. Personal hygiene 51 39 100 41 Food hygiene 41 31 85 35 Informants who were unable to re- Other preventive measures 7 5 12 5 call at least three preventive measures Total responses 131 99a 241 99a 0.544b from memory were 2.8 times as likely to self-report inconsistent use of the a Percentage for total responses adds to 99% because of rounding in the percentages for the general thematic classes. b Frequency of cholera preventive measures (grouped according to general thematic class) recalled by household informants measures in the household (Table 7). and household infection status was not statistically significant for nonrandom association by chi-square test for independence Similarly, informants who self-reported (P = 0.544, degrees of freedom = 3).

TABLE 6. Prevalence of parasitic infection according to informant’s ability to recall cholera preventive measures from memory, self- reported usage of cholera prevention measures, and source of household drinking water, Valle Hermoso, Tamaulipas, Mexico. 1994–1997

Household infection status Prevalence 95% confidence limit Variable Infected Uninfected odds ratio Lower Upper

Ability to recall three or more cholera prevention measures from memory Not able 38 42 2.018a 1.067 3.819 Able 26 58 Self-reported household use of cholera prevention measures Inconsistent 23 36 0.865b 0.430 1.792 Consistent 31 42 Source of drinking water used in household Untreated, used directly from faucet 30 12 5.111c 2.393 10.914 Bottled, chemically treated, or filtered 45 92 a The frequency of informants able to recall three or more cholera preventive measures and household infection status was statistically significant for nonrandom association by Fisher’s exact test (P = 0.037, degrees of freedom (df) = 1). b The frequency of informants’ self-reported household use of cholera preventive measures and household infection status was not statistically significant for nonrandom association by Fisher’s exact test (P= 0.724, df = 1). c The frequency of households using untreated water and household infection status was statistically significant for nonrandom association by Fisher’s exact test (P < 0.0001, df = 1).

Rev Panam Salud Publica/Pan Am J Public Health 14(1), 2003 37 TABLE 7. Self-reported usage of cholera preventive measures in household according to informant’s ability to recall preventive measures from memory and source of drinking water used in household, Valle Hermoso, Tamaulipas, Mexico. 1994–1997

Self-reported household use of cholera prevention measures 95% confidence limit Prevalence Inconsistent Consistent odds ratio Lower Upper

Ability to recall three or more cholera prevention measures from memory Not able 37 30 2.775a 1.359 5.666 Able 20 45 Source of drinking water used in household Untreated, used directly from faucet 20 10 3.657b 1.542 8.676 Bottled, chemically treated, or filtered 35 64 a The frequency of informants’ self-reported inconsistent use of cholera prevention measures and their inability to recall from memory three or more cholera prevention measures was statistically significant for nonrandom association by Fisher’s exact test (P = 0.0052, degrees of freedom (df) = 1). b The frequency of informants self-reporting inconsistent use of cholera prevention measures and their use of untreated water obtained directly from the faucet was statistically significant for nonrandom association by Fisher’s exact test (P = 0.0031, df = 1).

pend on the fecal-oral transmission low recommendations to engage in fre- risk of cholera infection is reflected in route (5). These conclusions were based quent hand-washing and other health- the lower frequency of other enteric on comparisons of the total number of promoting behaviors for prevention of pathogens that also depend on the parasitic infections diagnosed in San- infectious enteric diseases has not been fecal-oral transmission route. tiago area hospitals in the 3 years pre- the subject of systematic investigation The overall prevalence of endopara- ceding and the 3 years following a na- despite the widespread acknowledg- sitic infection in the Valle Hermoso tional cholera prevention campaign. ment that parasitic infections are evi- children was 30%. Parasites capable of However, interviews were not con- dence of fecal-oral contamination. producing disease accounted for 20% ducted with cases and noncases to de- Species such as Entamoeba coli, Enta- of the infections. Giardia lamblia, a fla- termine if these persons had actual moeba hartmanni, Iodamoeba bütschlii, gellated protozoan associated with di- knowledge of the ways to prevent and Endolimax nana are especially well arrheal disease, was the most prevalent cholera, or if the prevention methods suited to this role because they do not of the pathogenic species. Although G. were used in their households. The ab- cause clinical disease and are not tar- lamblia infection can be treated with sence of this information limits the geted for treatment by physicians or metronidazole and albendazole, which usefulness of the observation because community-based drug distribution are available as “over the counter” the association may be coincidental programs. Infections are indirectly fa- drugs in Mexican pharmacies, the par- and because the findings are poten- cilitated in susceptible hosts by con- asite is highly infectious and rapidly tially unrelated. sumption or preparation of food with spread from one infected person to an- In this study in Mexico we inves- unwashed hands following defecation other. Such treatment and control ef- tigated the hypothesis that children or changing an infant’s diaper. The oc- forts are frustrated by the parasite’s 5- residing in households where the currence of these nonpathogenic pro- day to 14-day lifecycle and its ability to primary caregiver had knowledge of tozoa in the feces of individuals shar- parasitize human and nonhuman hosts Government-promoted cholera pre- ing the same household is indirect without causing disease (13). Asymp- vention recommendations were less evidence of a range of behaviors nec- tomatic infection in these hosts is an likely to be infected with intestinal par- essary to maintain fecal-oral transmis- important reservoir for spread of infec- asites dependent on fecal-oral trans- sion of the parasites between infected tious cysts to susceptible persons. Ef- mission. Fecal samples from resident and susceptible persons. fective control of the parasite requires children were examined for parasitic The results of this investigation indi- a sustained effort to break the cycle of infection, and these data were used cate that children residing in house- infection and re-infection with effec- as biological indicators of household holds where the primary caregiver was tive chemotherapy and promotion of knowledge of and compliance with unable to recall from memory at least frequent hand-washing and other per- cholera prevention measures. The fecal- three cholera prevention measures sonal hygiene measures. oral nature of transmission for many were two times as likely to be infected The lower prevalence of infections intestinal protozoan and helminth para- with an intestinal parasite. Although with Ascaris lumbricoides and Trichuris sites that infect humans is well estab- we cannot posit a clearly defined cause- trichiura reflects the effect of suppres- lished (13). However, recognition that and-effect relationship between these sive targeted chemotherapy for school- nonpathogenic protozoan species can variables, the significance of the associ- age children. In Valle Hermoso, school- serve as biological indicators of how ation suggests that greater awareness age children (at least 5 years old) were well individuals and households fol- of the hygienic measures that reduce treated with albendazole (a single dose

38 Faulkner et al. • Prevalence of endoparasitic infection in children and cholera prevention efforts in Mexico of 400 milligrams) during the study Infected children were significantly rect” answer, that is, that preventive period as part of the Government’s older and almost three times as likely measures were employed in the house- parasite control program. However, to have an infected sibling as were hold on a consistent basis. However, in- the fecal sample collection activities for children from uninfected households. formants able to list at least three pre- our study were conducted at least 90 Households with three or more chil- ventive measures from memory were days following that treatment, so that dren were nearly five times as likely also more likely to report consistent the children had sufficient time to de- to have an infected child as compared use of the preventive measures in the velop mature infections with A. lumbri- to households with two or fewer chil- household. An association between coides and T. trichiura (13). It is well dren. It is intuitive that the risk of para- self-reported inconsistent use of the pre- established that the long-term effect of sitic infection in children should in- ventive measures in households and suppressive drug therapy for school- crease with age as they become more the use of untreated drinking water was age children should result in a net re- ambulatory and their school and play noted. Together, these observations duction in the number of infective eggs activities take them farther from the suggest that the ability to recall specific available in the environment for rein- watchful eye of their mothers and other preventive measures may be a valid fection of new hosts in the community caregivers. Infected siblings also pro- proxy for household use of those mea- (14). However, the persistence of T. tri- vide a reservoir and a mechanism for sures, and this is reflected by the lower chiura infections in school-age children introducing parasites that can be shared prevalence of parasitic infection. who should be benefiting from peri- by all family members in the household The impact of Mexico’s national odic administration of anthelmintics (16). The strained supervisory abilities cholera prevention campaign on the indicates continuing transmission of of caregivers in households with nu- occurrence of endoparasitic infection the parasite in the community and merous children is a plausible explana- in Valle Hermoso cannot be directly as- raises an important issue regarding the tion for the fact that increased family sessed in the absence of community- efficacy of these drug treatments for size is an exposure variable associated wide parasite prevalence data collected control of this parasite. The median with parasitic infection. prior to 1990. Statistical data compiled cure rate for T. trichiura is only 38%, al- In Valle Hermoso the majority of by state health departments in Mexico though individual studies report cure household informants were able to are of limited use because they are rates that range from 4.9% to 99.3% recall from memory at least one inter- based on cases of parasitic infection (15). This widespread variation in ef- vention for preventing cholera in- that require the attention of medical ficacy indicates that periodic surveil- fection. The ability to recall specific authorities and thus do not include lance is necessary to monitor the recommendations for treatment of nonpathogenic protozoan parasites success of anthelmintic treatment pro- drinking water, personal hygiene, and that are not associated with clinical dis- grams for school-age children, with an food hygiene was similar for infor- ease and treated. Communitywide par- eye toward the potential emergence of mants from uninfected and infected asite surveys conducted in the Mexican drug-resistant parasite populations. households. Personal hygiene mea- states of Morelos, Nuevo Leon, and Infections with nonpathogenic pro- sures, such as hand-washing before Sinaloa (19–23) are likewise of limited tozoan species such as E. coli, E. hart- meals and after using the bathroom, utility because of differences in study manni, E. nana, and I. bütschlii and the were listed with the highest frequency. objectives, community hygiene, stan- tapeworm Hymenolepis nana indicate Food hygiene measures, such as wash- dard of living, and access to potable the persistence of inadequate hygienic ing fruits and vegetables and cook- water. The relatively low (30%) preva- behavior in specific households. The ing foods completely, were second in lence of parasitic infection found in the prevalence of these species is probably frequency. Households infrequently samples examined during this study not affected noticeably by the semi- mentioned preventive measures re- indicates that many residents of this annual albendazole treatments used lated to seafood and consumption of community have benefited from public in the school chemotherapy program. food from street vendors. This was in- health education efforts promoted as a That is because the drug has limited teresting, given the important roles of response to the threat of epidemic efficacy against these parasites, and these two areas of risk in the epidemi- cholera. However, the 42% household the parasites are not treated by physi- ology of cholera and foodborne dis- prevalence of parasitic infection is cians because they are not associated ease outbreaks (17–18). troubling in light of the apparent ac- with clinical disease. The fecal-oral na- Informants from uninfected house- ceptance and adoption of the message ture of transmission and the relatively holds were significantly more likely to that “hygiene is health,” which is evi- short period of time required to es- be able to recall at least three preven- dent throughout the community, with tablish infections suggests that any tive measures. The lack of association it painted on walls in cafes, restaurants, chemotherapeutic benefit would be between household infection status and and public restrooms. Health and hy- short-lived in the absence of sustained self-reported use of cholera preventive giene educational posters illustrating personal hygiene measures required measures was not surprising. It likely the ways to prevent cholera and other to break the cycle of infection and re- reflects a bias toward providing inter- diarrheal diseases are similarly dis- infection of new hosts. view teams with the perceived “cor- played in neighborhood health clinics

Rev Panam Salud Publica/Pan Am J Public Health 14(1), 2003 39 and the small convenience stores fre- caregivers in households whose in- disease. Households in which a single quented by residents. fection status was unknown because child was examined are most suscepti- There was an association between they declined to submit fecal samples ble to misclassification as uninfected. household infection and informants for parasitologic analysis. Neither The greater likelihood of parasitic in- being unable to recall at least three the household characteristics nor the fection in households with at least cholera prevention measures. This in- range of knowledge and use of cholera three children may be confounded by dicates the persistence of the unhy- prevention measures provided by examination of a greater number of gienic behaviors that maintain fecal- these subjects differed appreciably fecal samples from those households, oral transmission of parasitic species from the study participants who did and it may not necessarily reflect in- between infected and susceptible provide a fecal sample. This finding creased probability of parasite trans- household members. By persisting in suggests that the data and results de- mission between infected and suscep- these risky behaviors, the residents of scribed in this paper are a reasonable tible family members. these households are at increased risk quick overview or “snapshot” of the Despite these limitations, this com- for morbidity associated with an out- health experience in the community at parative analysis of parasitic infection break of epidemic cholera as well as the time of the investigation. and its relation with cholera preven- other infectious parasitic and diarrhea- The cross-sectional study design tion efforts in this community provides related pathogens. limits the ability to ascertain whether a useful context for developing testable The conclusions drawn from this the onset of parasitic infection in the hypotheses for future investigations investigation are subject to several household is an antecedent or a con- based on cohort or case-control study limitations. The convenience sampling sequence of the exposure variables designs. Although costly in terms of methodology employed for collecting defined in this study. The prevalence money and labor, periodic surveillance fecal samples and interviewing house- odds ratio is the preferred measure for the nonpathogenic protozoan para- hold informants limits the ability to of association because the onset of dis- sites as bioindicators can be a useful generalize from this study to the rest ease is difficult to determine and be- method for evaluating the discrepancy of the residential population of Valle cause the duration of the outcome between informants who say they em- Hermoso and similar communities in (parasitic infection) is not affected by ploy recommended hygienic practices northeastern Mexico. Although this in- exposure status (12). Similarly, reli- for the prevention of diarrheal diseases vestigation was carried out in specific ance on a single fecal sample for dif- and informants who actually employ neighborhoods chosen for their coop- ferentiating infected and uninfected them consistently in their household eration with public health nurses and subjects may introduce a misclassifi- routines. The use of such studies as a local health promoters, this bias was cation bias. The sensitivity of a single method for monitoring compliance acceptable because the research inves- fecal sample is generally sufficient to with public health and hygiene pro- tigated the recall and use of cholera identify infected persons with symp- grams may indicate the need to inten- preventive measures in community toms of parasitic disease (24–25). Ex- sify educational efforts for prevention households. An important characteris- amination of multiple fecal samples of diarrhea associated with enteric tic for participation in the study was collected on successive days has been pathogens that cannot be controlled by the opportunity for exposure to Gov- shown to increase the sensitivity to drugs alone. ernment public health education pro- detect parasitic infections in asympto- grams. All participants were included matic persons (26). The prevalence of Acknowledgments. This research without regard to income, homeown- parasitic infection may be underrepre- was supported in part by a grant from ership, or family size. sented in this study because fecal sam- Sigma-Xi, The Scientific Research So- Interviews were also conducted ples were collected from healthy per- ciety (Research Triangle Park, North with a limited number of primary sons without symptoms of parasitic Carolina, United States).

REFERENCES

1. United States of America, Centers for Disease and control. Technical information from the ago poniente en relacion a la campaña de pre- Control. Cholera—Peru, 1991. MMWR Morb Communicable Diseases Program. Available vención del cólera. Parasitol al Día 1996;20: Mortal Wkly Rep. 1991;40(6):108–110. from: http://165.158.1.110/english/hcp/hct01. 27–31. 2. Jiménez Corona A, Gutiérrez Cogiot L, López htm [Internet site]. Accessed 29 May 2002. 6. México, Instituto Nacional de Estadística, Moreno S, Tapia Conyer R. El cólera en Mé- 4. México, Secretaría de Salud. Cólera en Mé- Geografía e Informática. Tamaulipas conteo xico: situación epidemiológica actual. Gac xico. Epidemiología Boletín Semanal 1994;1 de población y vivienda 1995. Resultados de- Med Mex 1995;131:363–366. (49):1–10. finitivos. Tabulados básicos. Aguascalientes, 3. Pan American Health Organization, Disease 5. Sotta P de la, Riquelme A, Silva B, Tassara R, México: Instituto Nacional de Estadística, Prevention and Control. Cholera prevention Muñoz P, López L. Enteroparasitosis en Santi- Geografía e Informática; 1996.

40 Faulkner et al. • Prevalence of endoparasitic infection in children and cholera prevention efforts in Mexico 7. México, Instituto Nacional de Estadística, geted chemotherapy through schools. Trans R los, Mexico. Am Jour Trop Med Hyg 1992;46: Geografía e Informática. Anuario estadístico Soc Trop Med Hyg 1990;84:115–120. 677–685. del estado Tamaulipas, edición 1996. Aguas- 15. Bennett A, Guyatt H. Reducing intestinal ne- 22. Vargas Mena J, Villarreal AC, Montes E. Fre- calientes: Instituto Nacional de Estadística, matode infection: efficacy of albendazole and cuencia de parásitos intestinales en el estado Geografía e Informática; 1996. mebendazole. Parasitol Today 2000;16:71–74. de Nuevo León, México: I, índices copropara- 8. Sloss MW, Kemp RL, Zajack AL. Veterinary 16. Mason PR, Patterson BA. Epidemiology of sitoscópicos en diez municipios del norte clinical parasitology. 6th ed. Ames, Iowa, Hymenolepis nana infections in primary school del estado. Rev Latinoam Microbiol 1970;12: United States of America: Iowa State Univer- children in urban and rural communities in 27– 33. sity Press; 1994. Zimbabwe. J Parasitol 1994;80:245–250. 23. Vargas Mena J, Rodríguez ME, Montes E. Fre- 9. Ash LR, Orihel TC. Atlas of human parasitol- 17. Albert JM, Neira M, Motarjemi Y. The role of cuencia de parásitos intestinales en el estado ogy. 3rd ed. Chicago, Illinois, United States: food in the epidemiology of cholera. World de Nuevo León, México: II, índices copro- American Society of Clinical Pathologists Health Stat Q 1997;50:111–118. parasitoscópicos en nueve municipios del ori- Press; 1990. 18. Moy G, Hazzard A, Käferstein F. Improving ente del estado. Rev Latinoam Microbiol 10. Garcia LS, Bruckner DA. Diagnostic medical the safety of street vended food. World Health 1970;12: 35–39. parasitology. 2nd ed. Washington, D.C: Amer- Stat Q 1997;50:124–131. 24. Montessori GA, Bischoff. Searching for para- ican Society for Microbiology; 1993. 19. Comacho SPD, Ruiz AC, Peraza VS, Ramos sites in stool: once is usually enough. Can 11. Price DL. Procedure manual for the diagnosis MLZ, Medina MF, Lozano R, et al. Epidemio- Med Assoc J 1987;137:702. of intestinal parasites. Boca Raton, Florida, logic study and control of Taenia solium infec- 25. Senay H, MacPherson D. Parasitology: diag- United States: CRC Press; 1994. tions with praziquantel in a rural village of nostic yield of stool examination. Can Med 12. Epidemiologic Research and Information Cen- Mexico. Am J Trop Med Hyg 1991;45:522–531. Assoc J 1989;140:1329–1331. ter. Cross-sectional studies. ERIC Notebook 20. Flores Castañeda M, Yáñez Rodríguez A, Tor- 26. Kawatu D, Lees RE, Maclachlan RA. Screen- 1999;(7):1–4. Available from: http:// www. res López E. Enteroparasitosis en una ing for intestinal parasites. Is a single sample sph.unc.edu/courses/eric/eric_notebooks. población estudiantil del área metropolitana valid? Can Fam Physician 1993;39:1748–1750. htm [Internet site]. Accessed 15 December 2002. de Monterrey, N.L. Boletín Epidemiológico 13. Beaver PC, Jung RC, Cupp EW. Clinical para- Fronterizo 1991;5:1–5. sitology. 9th ed. Philadelphia, Pennsylvania, 21. Sarti E, Schantz PM, Plancarte A, Wilson M, United States: Lea and Febiger; 1984. Gutierrez IO, Lopez AS. Prevalence and risk 14. Bundy DAP, Wong MS, Lewis LL, Horton J. factors for Taenia solium taeniasis and cysticer- Manuscript received 19 July 2002. Revised version ac- Control of geohelminths by delivery of tar- cosis in humans and pigs in a village in More- cepted for publication 21 February 2003.

RESUMEN Objetivos. Evaluar si un mejor conocimiento y una mayor aplicación de las medidas de salud pública fomentadas para la prevención del cólera se reflejan en una menor prevalencia de infestación parasitaria en el medio doméstico de una comunidad del es- Prevalencia de la infestación tado de Tamaulipas, México, cercano a la frontera con los Estados Unidos de América. endoparasitaria en niños Métodos. Entre 1994 y 1997 se recolectaron muestras fecales de 438 niños mediante un y su relación con muestreo por conveniencia. Estas muestras fueron examinadas en busca de huevos o lar- vas de helmintos y quistes de protozoos, como indicadores biológicos del cumplimiento los esfuerzos para en el medio doméstico de las medidas recomendadas para la prevención del cólera. Las la prevención del cólera medidas aconsejadas consistían en lavarse las manos antes de las comidas y después de en México defecar, tomar agua purificada, lavar las frutas y vegetales e ingerir alimentos bien coci- nados. Adicionalmente, se obtuvo información relativa al conocimiento y aplicación de las medidas para la prevención del cólera mediante entrevistas a adultos de 252 viviendas (186 de esas viviendas también entregaron muestras para análisis). Resultados. De los 438 niños, 131 (30%) presentaron infestaciones parasitarias. Estos niños residían en 79 (42%) de las 186 viviendas que entregaron muestras fecales. Gia- rdia lamblia representó el 12,5% de las infestaciones. También se encontraron Hy- menolepis nana, Ascaris lumbricoides, Trichuris trichiura, Enterobius vermicularis, Ancy- lostoma/Necator, Strongyloides stercoralis, Entamoeba coli, Entamoeba hartmanni, Entamoeba histolytica, Endolimax nana e Iodamoeba bütschlii. Los niños con parasitosis eran mayores y con mayor frecuencia tenían algún hermano o hermana infestado. Las viviendas con tres o más niños presentaron mayor probabilidad de tener algún niño infestado. Las personas encargadas de cuidar la salud en las viviendas donde había al menos un niño con parasitosis se caracterizaron por no poder mencionar de memoria al menos tres medidas de prevención contra el cólera. Conclusiones. La prevalencia de viviendas con niños infestados con parásitos (42%) fue relativamente elevada e indica que algunos residentes de esta comunidad pueden no haber respondido totalmente a los esfuerzos de educación sanitaria promovidos para la prevención del cólera. La presencia de protozoos parasitarios no patógenos, como Endolimax nana, Entamoeba coli, Entamoeba hartmanni o I. bütschlii, es un mar- cador biológico importante de la persistencia de hábitos higiénicos inadecuados que aumentan el riesgo de cólera y otras enfermedades infecciosas de transmisión fecal- oral. La información obtenida de estudios similares puede servir para vigilar el cumplimiento de los programas de salud e higiene comunitarias, e indica que es nece- sario intensificar el trabajo educativo dirigido a la prevención de la diarrea asociada con patógenos entéricos que no pueden ser controlados solo con medicamentos.

Rev Panam Salud Publica/Pan Am J Public Health 14(1), 2003 41