• View in gallery
    Figure 1.

    A, Interaction between baseline stunting and age for Raven score. The Raven’s colored matrices scores of children who were stunted (height-for-age Z-score [HAZ] < −2) at the beginning of the study did not vary with age (dotted line), while children who were not stunted at the beginning of the study demonstrated the expected correlation, with older children achieving higher scores (P = 0.025). B, Interaction between baseline HAZ and age for Raven score. C, Interaction between baseline weight-for-age Z-score [WAZ] and age for Raven score. Children with a higher baseline HAZ as well as higher baseline WAZ, showed a remarkably higher score change rate (P = 0.009 and 0.062, respectively). Three Z-score levels are shown for plot demonstration. These interaction models show that children who were malnourished before 3–5 years of age have difficulties four years later in attaining the same score gain as their counterparts who were not malnourished in early childhood. This may be due to some permanent damage to cognitive function from early childhood. However, the interactions may also be explained by the observation that most children who were malnourished upon entry into the study, as defined by baseline stunting, remained stunted throughout the duration of the study. Thus, malnutrition throughout childhood, rather than exclusively in early childhood, negatively influenced test scores.

  • 1

    Berkman DS, Lescano AG, Gilman RH, Lopez SL, Black MM, 2002. Effects of stunting, diarrhoeal disease, and parasitic infection during infancy on cognition in late childhood: a follow-up study. Lancet 359 :546–571.

    • Search Google Scholar
    • Export Citation
  • 2

    Niehaus MD, Moore SR, Patrick PD, Derr LL, Lorntz B, Lima AA, Guerrant RL, 2002. Early childhood diarrhea is associated with diminished cognitive function 4 to 7 years later in a northeast Brazilian shantytown. Am J Trop Med Hyg 66 :590–593.

    • Search Google Scholar
    • Export Citation
  • 3

    Haque R, Mondal D, Kirkpatrick BD, Akther S, Farr BM, Sack RB, Petri WA Jr, 2003. Epidemiological and clinical characteristics of acute diarrhea with emphasis on Entamoeba histolytica infections in preschool children in an urban slum of Dhaka, Bangladesh. Am J Trop Med Hyg 69 :389–405.

    • Search Google Scholar
    • Export Citation
  • 4

    Ayeh-Kumi PF, Ali IM, Lockhart LA, Gilchrist CA, Petri WAJ, Haque R, 2001. Entamoeba histolytica: genetic diversity of clinical isolates from Bangladesh as demonstrated by polymorphisms in the serine-rich gene. Exp Parasitol 99 :80–88.

    • Search Google Scholar
    • Export Citation
  • 5

    Haque R, Ali IM, Sack RB, Farr BM, Ramakrishnan G, Petri WAJ, 2001. Amebiasis and mucosal IgA antibody against the Entamoeba histolytica adherence lectin in Bangladeshi children. J Infect Dis 183 :1787–1793.

    • Search Google Scholar
    • Export Citation
  • 6

    Kordas K, Lopez P, Rosado JL, Garcia Vargas G, Alatorre Rico J, Ronquillo D, Cebrian ME, Stoltzfus RJ,2004. Blood lead, anemia, and short stature are independently associated with cognitive performance in Mexican school children. J Nutr 134 :363–371.

    • Search Google Scholar
    • Export Citation
  • 7

    Black MM, 2003. Micronutrient deficiencies and cognitive functioning. J Nutr 133 (11 Suppl 2):3927S–3931S.

  • 8

    Elbedour S, Bart WM, Hekter J, 2003. Intelligence and family marital structure: the case of adolescents from monogamous and polygamous families among Bedouin Arabs in Israel. J Soc Psychol 143 :95–110.

    • Search Google Scholar
    • Export Citation
  • 9

    Daley TC, Whaley SE, Sigman MD, Espinosa MP, Neumann C, 2003. IQ on the rise: the Flynn effect in rural Kenya. Psychol Sci 14 :215–219.

    • Search Google Scholar
    • Export Citation
  • 10

    Counter SA, Buchanan LH, Rosas HD, Ortega F, 1998. Neurocognitive effects of chronic lead intoxication in Andean children. J Neurol Sci 160 :47–53.

    • Search Google Scholar
    • Export Citation
  • 11

    Lai TJ, Guo YL, Guo NW, Hsu CC, 2001. Effect of prenatal exposure to polychorinated biphenyls on cognitive development in children: a longitudinal study in Taiwan. Br J Psychiatry 178 :S49–S52.

    • Search Google Scholar
    • Export Citation
  • 12

    Esquivel GB, 1965. Coloured progressive matrices. Buros OK, ed. Mental Measurements Yearbook. Highland Park, NJ: Gryphon Press, 206–212.

  • 13

    Raven J, Raven JC, Court JH, 1998. Section 2: Coloured Progressive Matrices. Manual for Raven’s Progressive Matrices and Vocabulary Scales. Oxford, United Kingdom: Oxford Psychologists Press, Ltd.

  • 14

    Hadidjaja P, Bonang E, Suyardi MA, Abidin SA, Ismid IS, Margono SS, 1998. The effect of intervention methods on nutritional status and cognitive function of primary school children infected with Ascaris lumbricoides. Am J Trop Med Hyg 59 :791–795.

    • Search Google Scholar
    • Export Citation
  • 15

    Gardner JM, Grantham-McGregor S, 1996. Baddeley A. Trichuris trichiura infection and cognitive function in Jamaican school children. Ann Trop Med Parasitol 90 :55–63.

    • Search Google Scholar
    • Export Citation
  • 16

    Checkley W, Epstein LD, Gilman RH, Black RE, Cabrera L, Sterling CR, 1998. Effects of Cryptosporidium parvum infection in Peruvian children: growth faltering and subsequent catch-up growth. Am J Epidemiol 148 :497–506.

    • Search Google Scholar
    • Export Citation
  • 17

    Guerrant RL, Schorling JB, McAuliffe JF, de Souza MA, 1992. Diarrhea as a cause and an effect of malnutrition: diarrhea prevents catch-up growth and malnutrition increases diarrhea frequency and duration. Am J Trop Med Hyg 47 :28–35.

    • Search Google Scholar
    • Export Citation
Past two years Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 234 96 10
PDF Downloads 92 48 0
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

COGNITIVE EFFECTS OF DIARRHEA, MALNUTRITION, AND ENTAMOEBA HISTOLYTICA INFECTION ON SCHOOL AGE CHILDREN IN DHAKA, BANGLADESH

JESSICA L. TARLETONDivision of Infectious Diseases and International Health and Division of Biostatistics and Epidemiology, University of Virginia, Charlottesville Virginia; International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh

Search for other papers by JESSICA L. TARLETON in
Current site
Google Scholar
PubMed
Close
,
RASHIDUL HAQUEDivision of Infectious Diseases and International Health and Division of Biostatistics and Epidemiology, University of Virginia, Charlottesville Virginia; International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh

Search for other papers by RASHIDUL HAQUE in
Current site
Google Scholar
PubMed
Close
,
DINESH MONDALDivision of Infectious Diseases and International Health and Division of Biostatistics and Epidemiology, University of Virginia, Charlottesville Virginia; International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh

Search for other papers by DINESH MONDAL in
Current site
Google Scholar
PubMed
Close
,
JIANFEN SHUDivision of Infectious Diseases and International Health and Division of Biostatistics and Epidemiology, University of Virginia, Charlottesville Virginia; International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh

Search for other papers by JIANFEN SHU in
Current site
Google Scholar
PubMed
Close
,
BARRY M. FARRDivision of Infectious Diseases and International Health and Division of Biostatistics and Epidemiology, University of Virginia, Charlottesville Virginia; International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh

Search for other papers by BARRY M. FARR in
Current site
Google Scholar
PubMed
Close
, and
WILLIAM A. PETRI JR.Division of Infectious Diseases and International Health and Division of Biostatistics and Epidemiology, University of Virginia, Charlottesville Virginia; International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh

Search for other papers by WILLIAM A. PETRI JR. in
Current site
Google Scholar
PubMed
Close

Cognitive function was assessed in 191 Bangladeshi children 6–9 years of age using verbal and nonverbal tests. These scores were added to a health surveillance database that was compiled over the four previous years that includes incidence of diarrhea and Entamoeba histolytica infection and nutritional status. The associations of diarrhea, malnutrition, and social factors with cognitive scores were analyzed statistically, and associations between diarrhea and test scores were controlled for the influence of social factors. Cognitive scores were negatively associated with stunting during school age, as well as the height-for-age and weight-for-age scores at study enrollment. Incidence of diarrhea was associated with nonverbal test scores before, but not after, controlling for socioeconomic factors. Generally E. histolytica infection was not found to independently influence scores, except that E. histolytica-associated dysentery was associated with lower test scores while dysentery of any etiology was not. Thus, malnutrition during the school age years, but not diarrhea or E. histolytica infection, was associated with a lower level of cognitive functioning. This suggested that intervention during school age years may be able to mitigate the cognitive deficiencies associated with malnutrition.

INTRODUCTION

Within the past several years, important research has linked early childhood diarrhea, which is diarrhea during the first two years of life, with deficits in cognitive abilities several years later.1,2 This association is attributed to the concept that the first two years of life represent a critical period in brain development, and that if severe diarrhea causes its characteristic dehydration and nutrient loss during this time, permanent or semi-permanent effects on cognitive function may result. Although these studies associate illness in the first two years with cognitive function later in life, they have not reported the influence of disease, malnutrition, and social environment in the years after infancy and before the cognitive assessment. Consequently, it is not known whether diarrhea in the first two years of life exclusively causes the cognitive differences seen years later, or if higher levels of diarrhea during early childhood simply foreshadow illness and environmental risk factors throughout childhood.

In addition to the timing of diarrheal illness, different etiologies of diarrhea may result in unique physical effects according to their distinct pathogeneses. Entamoeba histolytica is one of the many intestinal parasites that causes diarrhea in impoverished areas. In places such as Dhaka, Bangladesh, where the current study was carried out, E. histolytica is associated with 8% of diarrheal episodes and 8.7% of dysenteric episodes.3 However, an estimated 90% of all patients colonized with E. histolytica never manifest clinical symptoms.4 This distinctive pattern of illness calls for a more complete understanding of the consequences of E. histolytica infection.

The current study sought to address these gaps in knowledge by assessing the cognitive impact of diarrheal disease and E. histolytica infection in children 2–9 years of age, ages for which the effect of diarrheal illness on cognitive performance has not yet been reported. Towards these ends, the current study undertook both verbal and nonverbal cognitive testing of a cohort of 191 children 6–9 years of age living in an urban slum of Dhaka, the capital of Bangladesh. Comprehensive health surveillance of these children had taken place in the four years before cognitive testing beginning in these children’s second to fifth year of life. Cognitive scores were correlated with data on diarrheal illness from the preceding four years, including the frequency, duration, and etiology of all diarrheal episodes; the amount of time colonized with E. histolytica and its associated diarrheal and dysenteric episodes; as well as anthropometric data such as height and weight to evaluate nutritional status. Additionally, socioeconomic and educational history surveys for each child were completed to estimate the effects of the social environment on cognitive scores.

METHODS

Enrollment and health surveillance.

The study population was composed of 191 children from Mirpur, an urban slum in Dhaka, Bangladesh. Enrollment and health surveillance took place as described elsewhere.3,5 Children and their parents were visited and interviewed every other day by health care workers and diarrheal stools were tested for E. histolytica over a three-year period. In addition, surveillance stool specimens were obtained every month for detection of E. histolytica infection. All enrolled children and their family members received free primary health care services, including medications, from the project office in Mirpur. Episodes of diarrhea were treated with oral rehydration and antibiotics or anti-amebic medications as needed. Informed consent was obtained from the parents or guardians. The human experimentation guidelines of the U.S. Department of Health and Human Services, the University of Virginia, and the Center for Health and Population Research (ICDDR,B) (Dhaka, Bangladesh) were followed in this research.

Further demographic information was gathered at the time of cognitive testing, including level of parental education, start date and duration of schooling for each child, type of school attended, family size and number of siblings, participation in drawing classes, music classes, sports, or part-time or full-time work, and family possession of a television or radio. These data were examined for statistically significant associations with cognitive abilities, as were environmental variables independent of health that could influence cognitive scores. Blood ferritin level was assessed during the summer of 2003 because low iron levels have been associated with cognitive performance.6,7

Diagnosis of E. histolytica infection.

Entamoeba histolytica infection was diagnosed by detection of antigen in stool using the E. histolytica II test and the morphologically identical E. dispar infection by the Entamoeba stool antigen detection test (both obtained from TechLab, Inc., Blacksburg, VA).

Test selection and administration.

The colored progressive matrices (CPM) test was selected for its non-reliance on language, worldwide acceptability throughout disparate cultures, and ease of use. The CPM test is one of a series of Raven’s matrices, including the standard progressive matrices and the advanced progressive matrices that have been used worldwide as reliable non-verbal measures of general intelligence.811 The CPM test has shown high reliability for young children, non-English speakers, and across different cultures.12 Test administration was followed according to the CPM manual.13 Instructions were explained to the children in Bangla, and one practice question was administered to ensure understanding.13

The definitions section of the Wechsler abbreviated scale of intelligence (WASI) was selected as a verbal measure of cognitive abilities, and adapted as described in this report to fit the characteristics of the study population. The definitions section is one of four subtests, two verbal and two nonverbal, of the WASI. The original WASI, which was standardized on a large sample of English-speaking residents of the United States, is intended to be a quick test of general intelligence. The test is composed of 38 items, with the first 4 requesting that the child give the name of a picture that is shown to him or her, and the remaining 34 requesting that a child verbally define words provided orally. The definitions section of the WASI test was translated into Bangla, the official language of Bangladesh, by two bilingual English-Bangla researchers in the parasitology laboratory of the ICDDR,B. In addition, several cultural adaptations were made with the consultation of Bengali researchers of the parasitology laboratory, as well as a psychologist on site. Each question was scored with from 0 to 3 points using the criteria given in the WASI manual.

Since the CPM test and the WASI were not standardized for the subject population, and changes were made to the definitions subtest that was used, these scores could not be translated into IQ or compared with other children external to the study. Raw scores were used to compare between the subject children. Scores were adjusted for age and months of schooling in statistical analysis.

Three field health workers were trained to administer the CPM test and the definitions test of the WASI. One hundred ninety-one children were given the CPM test in June 2003 over a three-week period. One hundred eighty-eight children took the definitions test, which was administered from early July through August 2003 by the same three health workers. Three children did not sufficiently demonstrate understanding of the definitions test according to the administration instructions. These tests were scored by the authors. The children were randomly assigned to one of the three health workers who administered the test. There were no statistically significant differences in mean test scores for tests administered by the three different health workers.

Statistical analysis.

Associations between cognitive scores (CPM test and WASI) and predicting variables were examined through both simple regression and multiple regression. Simple regression was performed to determine the influence of illness, malnutrition, and socioeconomic conditions on cognitive scores. Multiple regression was used to control for the influence of socioeconomic conditions on the relationships between illness and malnutrition, and the test scores. Socioeconomic condition variables examined for influence on test scores were age, months of school, type of school, sex, mother’s and father’s age and level of education, number of family members, number of siblings, income, possession of a television, possession of a radio, number of times held back in school, participation in sports, drawing and music classes, and having a job. Additional health-related variables examined were baseline and end-point ferritin levels, percent time positive for Ascaris lumbricoides and Trichuris trichiura, and immunization level. The adjusting socioeconomic condition variables were selected by automatic stepwise selection with both entry and exit P values = 0.15. The selection result was manually checked and only variables with a P value ≤ 0.05 were included in the final selected adjustment group. For all analyses, regression parameter estimates (betas) and their 95% confidence intervals were reported. Beta estimate indicates the slope of the regression line, which is equivalent to the outcome change when the continuous predictor changes a unit, or in the case of a categoric variable, the difference from the reference group. For those relationships between outcome variables and predicting variables that are not simple linear and could not be corrected by typical transformations, the variables were dichotomized based on interpretation and distribution.

RESULTS

Characteristics of the population.

The characteristics of the study population are shown in Table 1. Of 181 families for which income was known, the average income was 4,200 takas (approximately US$68) per month. Only 25% of the fathers and 15% of the mothers had more than five years of formal education.

From anthropometric data (height, weight, upper arm circumference) gathered both at the beginning of the study and every four months thereafter, 29% of children were stunted (height-for-age Z-score [HAZ] < −2) upon entry to the study, and 34% were underweight at baseline (weight-for-age Z-score [WAZ] < −2). Most of the children stunted at study enrollment (64%) remained stunted throughout the four years of the study.

Over the complete four-year study period, 191 children experienced 2,592 days of diarrhea (13.6 per child) in 1,269 total episodes. Sixty-seven percent of the children had less than eight episodes of diarrhea during the study period. Eight percent of all diarrheal stool samples available for diagnosis (1,269) were associated with E. histolytica infection. Nineteen children also experienced E. histolytica-associated dysentery during this time. Despite the fact that approximately 10% of the children experienced E. histolytica-associated diarrhea, 79% of the children were found to be asymptomatically colonized with E. histolytica at some point during the study, with 34 children carrying asymptomatic E. histolytica more than 10% of the study period. Eight-six percent of children also tested positive for other intestinal parasites during 50% or more of the study period.

Simple regression.

Environmental and socioeconomic variables that significantly predicted test scores (P < 0.05) are shown in Tables 2 and 3. Both Raven’s colored progressive matrices and WASI definitions scores were significantly associated with age, mothers’ education, and fathers’ education, each with P < 0.0005. Children with household incomes of ≤ 5,000 takas (approximately US$81) had significantly lower CPM scores than children with household income of 5,000–7,000 takas (P = 0.0378) or > 7,000 takas (P = 0.0042). Verbal definitions scores only differed significantly between families that earned ≤ 5,000 takas and those that earned > 7,000 takas per year (P = 0.0141). The CPM scores of children who attended non-governmental organization (NGO)–sponsored schools did not differ from those who did not attend school, but children who attend public school showed significantly higher CPM scores (P = 0.0032 between no school and public school; P = 0.0062 between NGO-sponsored school and public school). A similar trend was apparent in verbal definitions scores, which differed significantly only between no school and public school (P = 0.0335).

Malnutrition and diarrhea variables that were significantly associated with test scores are shown in Tables 4 and 5. Baseline stunting (HAZ ≤ −2) was associated with low scores on both the CPM and verbal definitions tests (P = 0.005 and P = 0.001, respectively). This association was seen even stronger when baseline HAZ was examined as a continuous variable, with P < 0.0001 for each test. Similarly, lower baseline WAZ scores showed significant associations with lower CPM and verbal definitions scores (P = 0.0023 and P = 0.0008, respectively). Furthermore, the percentage of study time that a child was stunted was positively associated with both CPM and verbal definitions scores (P = 0.0070 and P < 0.0001, respectively). In addition, the percent of the study time a child had other intestinal parasites (A. lumbricoides and T. trichiura) was a significant predictor of both CPM and verbal test scores (P = 0.036 and P = 0.0045, respectively). The CPM scores were significantly lower in children who had diarrhea more than 1% of the study period (P = 0.0439), as well as in children who had eight or more episodes of diarrhea during the study period (P = 0.0226). In contrast, verbal definitions scores were not significantly associated with episodes of diarrhea.

Although E. histolytica-associated diarrhea, percentage time colonized with E. histolytica, and asymptomatic E. histolytica infection did not significantly influence either CPM or verbal scores, a borderline correlation did emerge between the number of episodes of E. histolytica-related dysentery and verbal, but not CPM, test scores (P = 0.057). Nineteen of the 191 children had 24 E. histolytica-related episodes of dysentery. In contrast, there was no significant association found between dysentery of all etiologies and test scores either before or after controlling for environmental variables.

Multiple regression with adjustment for socioeconomic variables.

Raw scores from the Raven’s and WASI tests were controlled for age, months of schooling, and mother’s level of education, and the WASI definitions score was additionally controlled for family possession of a radio (Tables 4 and 5). Adjusting for mothers’ education completely accounted for the level of fathers’ education and all other significant socioeconomic variables.

Interestingly, interactions existed between the children’s age and baseline stunting, baseline HAZ score and baseline WAZ score for CPM score (Figure 1), but they were not significant for the verbal definition score. The interaction models indicated that CPM scores increased depending on the age of the children but that this increase differed based on the children’s baseline nutrition status. This difference is most clearly demonstrated in Figure 1A: older children who were stunted at baseline demonstrated very little difference in score with younger children stunted at baseline, while those without baseline stunting showed significant difference based on age. Similarly, children with lower baseline HAZ or WAZ scores showed significantly slower rate of change of CPM scores with age increase (Figure 1B and C). The values of −1, 0, and 1 for both HAZ and WAZ were chosen arbitrarily in these models to demonstrate trends between children having these Z scores. The association between the verbal definitions test and baseline stunting and baseline WAZ score was of only borderline significance (P = 0.084, and P = 0.089, respectively) while the association with baseline HAZ score remained strong under multiple regression (P = 0.008) (Table 5).

There were no significant associations of either the CPM or the verbal definitions tests with percentage of days with diarrhea or number of episodes during the study period by multivariate analysis (Tables 4 and 5). However, a significant association remained between E. histolytica-related dysentery and the verbal definitions test (P = 0.050). Additionally, the relationship between the percentage of the study time infected with A. lumbricoides or T. trichiura was significant after adjusting for socioeconomic variables (P = 0.023).

DISCUSSION

These data show that children who were stunted throughout the study period performed significantly worse on the WASI definitions test after adjustment for environmental factors. Previous studies have proposed that because diarrheal disease and malnutrition in early childhood are significantly associated with cognitive test scores years later, that diarrhea and malnutrition between 0 and 2 years of age are critically and irreversibly damaging to lifetime cognitive development. In contrast, this study demonstrates that malnutrition after the age of two is significantly associated with one measure of verbal skills, indicating that malnutrition later in childhood may contribute to cognitive deficits. Interventions during later childhood might be able to ameliorate some of the impairment observed, and further studies should test the possible importance of nutritional supplementation after age two in cognitive development.

Also notable are the interaction models between CPM scores, age, and baseline nutritional status, which indicate that children who were malnourished before ages 2–5 have difficulties four years later in attaining the same scores as their counterparts who were not malnourished in early childhood. This may be due to some permanent damage to cognitive function from early childhood as previously postulated. However, very instructive in interpreting all our data is the finding that most children who were stunted at the start of the study remained stunted for the four subsequent years. Thus, malnutrition throughout childhood, rather than exclusively in early childhood, negatively influenced test scores. Although it is impossible to attribute deficiencies in test score to any one period of childhood without data before the children’s enrollment, and this study could not establish causation of lower cognitive scores by any of the nutritional, diarrheal, or social factors considered, the associations presented stress that later childhood cannot be ignored in future research and public health interventions.

The study was designed to obtain a broad assessment of general intelligence by using both verbal and nonverbal measures. Although both Raven’s colored progressive matrices and WASI are used as measures of general intelligence, distinct patterns were observed in the effects of malnutrition and diarrhea on the different assessments. These differences may be the result of both the verbal versus nonverbal nature of the two tests, or because the definitions test of WASI is normally used as only one subtest of four used to gain a full assessment of general intelligence. The most interesting difference is that seen between test scores and the duration of stunting during the four-year data collection period: although both CPM and definitions scores were significantly associated with baseline stunting, children who were stunted throughout the four-year study period showed a very significant shortfall in verbal test scores compared with those who were stunted less than 100% of the four-year period (P = 0.006). They did not demonstrate such a deficit in nonverbal test scores (P = 0.237). This would seem to indicate that poor nutrition during the ages of 2–9 years is more influential in a child’s language development than in more general measures of nonverbal intelligence, such as the CPM test. Testing of more specific abilities is needed to validate this observation.

Childhood diarrhea between the ages of 2 and 9 years was not a statistically significant predictor of test scores independently of environmental and socioeconomic variables. Furthermore, E. histolytica infection and its related diarrhea do not independently affect CPM or WASI scores. The disappearance of the significant association between CPM or verbal definitions tests and percentage of days with diarrhea or number of episodes during the study period under the multiple regression analysis can be at least partially explained by the fact that age is highly correlated with percentage of days with diarrhea and number of episodes during the study period. Older children had highly significantly lower percentage of days with diarrhea and lower number of episodes during the study period. Therefore, when age was included in the models, the association does not remain statistically significant. The percentage of the study time infected with other intestinal parasites (A. lumbricoides or T. trichiura) was a significant predictor of verbal test scores after adjusting for socioeconomic variables. The resolution of A. lumbricoides infection has been previously shown to improve cognitive function,14 while the resolution of T. trichiura does not.15

Children with E. histolytica-associated dysentery performed significantly worse on the WASI definitions test after adjustment for environmental factors (P = 0.05). This association was not observed for dysentery of all etiologies. However, children with E. histolytica-associated dysentery (19 of 191) may have experienced considerably more frequent episodes of diarrhea than the rest of the study population, so that E. histolytica-associated dysentery serves as a marker of children with very severe diarrhea and parasitic infection. For example, children who experienced E. histolytica-associated dysentery had an average of 11.2 episodes of diarrhea in four years, which put them in the top 15% of number of episodes of diarrhea among all children.

Future studies would greatly benefit from enrolling subjects from birth and surveying both malnutrition and disease starting from birth and continuing through late childhood. In addition, breastfeeding seems to have a strong impact on early childhood development as well as protection against disease, and would be an interesting addition to comprehensive studies of childhood diarrhea and cognitive function.

Furthermore, studies from birth should record the timing of stunting and incidence of diarrhea throughout childhood to further define critical periods, if any, during childhood. Several studies have observed that the timing of parasitic infection determines if a child will achieve normal height later in life (known as catch-up growth).16 This finding is supported by the finding that diarrhea prevents catch-up growth of children malnourished in childhood.17 Could diarrhea also prevent “cognitive catch-up”? A more profound understanding of the influence of the timing of malnutrition and diarrheal diseases on cognitive function will require a longer-term study than has yet been realized.

Table 1

Demographic characteristics of the study population*

Mean SD
Age (months) 102.7 11.7
No. %
* NGO = non-governmental organization school, i.e., religious-based school (non-public school).
† One Taka = US $62.
Type of school (n = 186)
    No school 34 18.28
    Primary school 77 41.4
    NGO school 75 40.32
Sex (n = 191)
    Male 92 48.17
    Female 99 51.83
Father’s level of education (n = 191)
    No formal 95 49.74
    Primary 46 24.08
    Mid-level and higher 50 26.18
Mother’s level of education (n = 191)
    No formal 123 64.4
    Primary 47 24.61
    Mid-level 21 10.99
Family income, takas† (n = 180)
    ≤ 5,000 137 76.11
    5,000–7,000 21 11.67
    > 7,000 22 12.22
Household possession of a radio (n = 190)
    No 109 57.37
    Yes 81 42.63
Household possession of a television (n = 188)
    No 47 25
    Yes 141 75
Months of schooling (n = 191)
    0 34 17.80
    1–6 54 28.27
    7–18 38 19.90
    19–30 31 16.23
    30–42 16 8.38
    ≥ 43 18 9.42
Table 2

Variables significantly predicting results of the colored progressive matrices (CPM) test in simple regression*

95% Confidence interval
Predicting variable Beta estimate† Lower Upper P
* NGO = non-governmental organization.
† Beta estimate indicates the slope of the regression line such that a beta estimate of 0.10 for the relationship between age of child and CPM test scores indicates that with each month older a child was he or she scored 0.1 points better on the CPM test. This applies also to Tables 3, 4, and 5.
Age of child (months) 0.10 0.05 0.15 0.0001
Months of schooling 0.09 0.06 0.12 < 0.0001
Type of school None Reference
Primary 2.52 0.86 4.19 0.003
NGO 0.68 −1.00 2.35 0.426
Mothers’ level of education None Reference
Primary level 0.64 −0.76 2.04 0.372
Mid-level 3.57 1.64 5.49 0.0003
Fathers’ level of education None Reference
Primary level 0.87 −0.60 2.34 0.247
Mid-level 2.50 1.06 3.93 0.001
Family income, Takas < 5,000 Reference
5,000–7,000 2.07 0.12 4.03 0.038
> 7,000 2.82 0.90 4.73 0.004
Family possession of radio No Reference
Yes 1.51 0.29 2.73 0.016
Table 3

Variables significantly predicting results of the WASI in simple regression*

95% Confidence interval
Predicting variable Beta estimate Lower Upper P
*WASI = Wechsler abbreviated scales of intelligence; NGO = non-governmental organization.
Age of child (months) 0.15 0.09 0.22 < 0.0001
Months of schooling 0.13 0.09 0.17 < 0.0001
Type of school
    None Reference
    Primary 2.31 0.18 4.43 0.034
    NGO 0.88 −1.25 3.01 0.417
Family income, Takas
    < 5,000 Reference
    5,000–7,000 1.43 −1.11 3.98 0.267
    > 7,000 3.20 0.65 5.74 0.014
Mothers’ level of education
    None Reference
    Primary 1.97 0.14 3.79 0.035
    Middle 4.62 2.06 7.19 0.0005
Family possession or radio
    No Reference
    Yes 2.93 1.35 4.50 0.0003
Currently taking art class in school
    No Reference
    Yes 2.86 0.57 5.15 0.015
Table 4

Significant associations between Raven’s colored matrices scores and malnutrition and diarrhea

Unadjusted (simple regression) 95% Confidence interval Adjusted for predicting demographic data* (multiple regression) 95% confidence interval
Variables Beta estimate Lower Upper P Beta estimate Lower Upper P
* Adjusted for age, mother’s level of education, and months of schooling.
† HAZ = height-for-age Z-score as defined by the World Health Organization.
‡ WAZ = weight-for-age Z-score as defined by the World Health Organization.
§ Ascaris lumbricoides and Trichuris trichiura.
Baseline stunting HAZ† > −2 Reference
HAZ ≤ −2 −1.90 −3.21 −0.58 0.005 11.03 0.63 21.42 0.038
Baseline HAZ 0.94 0.54 1.34 < 0.0001 −4.12 −7.64 −0.59 0.022
Baseline HAZ/age interaction 0.05 0.01 0.08 0.009
Baseline WAZ‡ 0.95 0.35 1.56 0.002 −3.96 −8.64 0.73 0.097
Baseline WAZ/age interaction 0.04 −0.002 0.09 0.062
Stunting duration (continuous) −0.02 −0.03 −0.01 0.007 −0.01 −0.02 0.01 0.291
Percent of study period with other intestinal parasites§ < 30% Reference
≥ 30% −3.43 −6.64 −0.22 0.036 −2.49 −5.42 0.44 0.096
Percentage of study period with diarrhea ≤ 1.0% Reference
> 1.0% −1.40 −2.77 −0.04 0.044 −0.79 −2.06 0.49 0.226
Diarrheal episodes 0–7 episodes Reference
8+ episodes −1.49 −2.77 −0.21 0.023 −0.69 −1.92 0.53 0.266
Table 5

Significant associations between Wechsler abbreviated scales of intelligence definitions scores and malnutrition and diarrhea*

Unadjusted (simple regression) 95% Confidence interval Adjusted for predicting demographic data† (multiple regression) 95% confidence interval
Variables Beta estimate (95% confidence interval) Lower Upper P Beta estimate Lower Upper P
* For definitions of abbreviations, see Table 4.
† Adjusted for age, mother’s level of education, family possession of radio, and months of schooling.
Baseline stunting WAZ > −2 Reference
WAZ ≤ −2 −2.93 −4.63 −1.23 0.001 −1.42 −3.03 0.19 0.084
Baseline HAZ 1.20 0.67 1.73 < 0.0001 0.66 0.14 1.18 0.013
Baseline WAZ 1.37 0.58 2.17 0.001 0.65 −0.10 1.40 0.089
Stunting duration (continuous) −0.04 −0.06 −0.02 < 0.0001 −0.02 −0.04 −0.01 0.008
Percent of study period with < 30% Reference
intestinal parasites ≥ 30% −6.04 −10.19 −1.89 0.0045 −4.29 −7.97 −0.61 0.023
Diarrheal episodes 0–7 Episodes Reference
≥ 8 Episodes −1.47 −3.16 0.23 0.089 0.20 −1.37 1.78 0.798
Entamoeba histolytica-associated dysenteric episodes −1.84 −3.74 0.06 0.057 −1.68 −3.35 0.00 0.050
Dysenteric episodes of any etiology −0.16 −1.30 0.97 0.776 −0.001 −1.003 1.001 0.998
Figure 1.
Figure 1.

A, Interaction between baseline stunting and age for Raven score. The Raven’s colored matrices scores of children who were stunted (height-for-age Z-score [HAZ] < −2) at the beginning of the study did not vary with age (dotted line), while children who were not stunted at the beginning of the study demonstrated the expected correlation, with older children achieving higher scores (P = 0.025). B, Interaction between baseline HAZ and age for Raven score. C, Interaction between baseline weight-for-age Z-score [WAZ] and age for Raven score. Children with a higher baseline HAZ as well as higher baseline WAZ, showed a remarkably higher score change rate (P = 0.009 and 0.062, respectively). Three Z-score levels are shown for plot demonstration. These interaction models show that children who were malnourished before 3–5 years of age have difficulties four years later in attaining the same score gain as their counterparts who were not malnourished in early childhood. This may be due to some permanent damage to cognitive function from early childhood. However, the interactions may also be explained by the observation that most children who were malnourished upon entry into the study, as defined by baseline stunting, remained stunted throughout the duration of the study. Thus, malnutrition throughout childhood, rather than exclusively in early childhood, negatively influenced test scores.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 74, 3; 10.4269/ajtmh.2006.74.475

*

Address correspondence to William A. Petri Jr., Division of Infectious Diseases and International Health, University of Virginia Health System, Room 2115 MR4 Building, P.O. Box 801340, Charlottesville, VA 22908-1340. E-mail: wap3g@virginia.edu

Authors’ addresses: Jessica L. Tarleton and William A. Petri Jr., Division of Infectious Diseases and International Health, University of Virginia Health System, Room 2115 MR4 Building, P.O. Box 801340, Charlottesville, VA 22908-1340. Rashidul Haque and Dinesh Mondal, International Centre for Diarrhoeal Disease Research, Dhaka Bangladesh. Jianfen Shu and Barry M. Farr, Division of Biostatistics and Epidemiology, University of Virginia, Charlottesville, VA 22908-1340.

Acknowledgments: We thank the parents and children of Mirpur for their participation and the field team, including field assistants and laboratory staff, for their involvement in the study.

Financial support: This study was supported by a grant (AI-43596) from the National Institutes of Health (NIH) and by the ICDDR,B. The ICDDR,B is supported by countries and agencies that share its concern for the health problems of developing countries.

Disclosure: William A. Petri Jr. received research support from TechLab, Inc. and royalties from a patent license agreement with TechLab for a diagnostic test for amebiasis. These royalties accrue to the American Society of Tropical Medicine and Hygiene without benefit to Dr. Petri. The NIH had no role in the design of the study but did review and approve the human investigation protocol.

REFERENCES

  • 1

    Berkman DS, Lescano AG, Gilman RH, Lopez SL, Black MM, 2002. Effects of stunting, diarrhoeal disease, and parasitic infection during infancy on cognition in late childhood: a follow-up study. Lancet 359 :546–571.

    • Search Google Scholar
    • Export Citation
  • 2

    Niehaus MD, Moore SR, Patrick PD, Derr LL, Lorntz B, Lima AA, Guerrant RL, 2002. Early childhood diarrhea is associated with diminished cognitive function 4 to 7 years later in a northeast Brazilian shantytown. Am J Trop Med Hyg 66 :590–593.

    • Search Google Scholar
    • Export Citation
  • 3

    Haque R, Mondal D, Kirkpatrick BD, Akther S, Farr BM, Sack RB, Petri WA Jr, 2003. Epidemiological and clinical characteristics of acute diarrhea with emphasis on Entamoeba histolytica infections in preschool children in an urban slum of Dhaka, Bangladesh. Am J Trop Med Hyg 69 :389–405.

    • Search Google Scholar
    • Export Citation
  • 4

    Ayeh-Kumi PF, Ali IM, Lockhart LA, Gilchrist CA, Petri WAJ, Haque R, 2001. Entamoeba histolytica: genetic diversity of clinical isolates from Bangladesh as demonstrated by polymorphisms in the serine-rich gene. Exp Parasitol 99 :80–88.

    • Search Google Scholar
    • Export Citation
  • 5

    Haque R, Ali IM, Sack RB, Farr BM, Ramakrishnan G, Petri WAJ, 2001. Amebiasis and mucosal IgA antibody against the Entamoeba histolytica adherence lectin in Bangladeshi children. J Infect Dis 183 :1787–1793.

    • Search Google Scholar
    • Export Citation
  • 6

    Kordas K, Lopez P, Rosado JL, Garcia Vargas G, Alatorre Rico J, Ronquillo D, Cebrian ME, Stoltzfus RJ,2004. Blood lead, anemia, and short stature are independently associated with cognitive performance in Mexican school children. J Nutr 134 :363–371.

    • Search Google Scholar
    • Export Citation
  • 7

    Black MM, 2003. Micronutrient deficiencies and cognitive functioning. J Nutr 133 (11 Suppl 2):3927S–3931S.

  • 8

    Elbedour S, Bart WM, Hekter J, 2003. Intelligence and family marital structure: the case of adolescents from monogamous and polygamous families among Bedouin Arabs in Israel. J Soc Psychol 143 :95–110.

    • Search Google Scholar
    • Export Citation
  • 9

    Daley TC, Whaley SE, Sigman MD, Espinosa MP, Neumann C, 2003. IQ on the rise: the Flynn effect in rural Kenya. Psychol Sci 14 :215–219.

    • Search Google Scholar
    • Export Citation
  • 10

    Counter SA, Buchanan LH, Rosas HD, Ortega F, 1998. Neurocognitive effects of chronic lead intoxication in Andean children. J Neurol Sci 160 :47–53.

    • Search Google Scholar
    • Export Citation
  • 11

    Lai TJ, Guo YL, Guo NW, Hsu CC, 2001. Effect of prenatal exposure to polychorinated biphenyls on cognitive development in children: a longitudinal study in Taiwan. Br J Psychiatry 178 :S49–S52.

    • Search Google Scholar
    • Export Citation
  • 12

    Esquivel GB, 1965. Coloured progressive matrices. Buros OK, ed. Mental Measurements Yearbook. Highland Park, NJ: Gryphon Press, 206–212.

  • 13

    Raven J, Raven JC, Court JH, 1998. Section 2: Coloured Progressive Matrices. Manual for Raven’s Progressive Matrices and Vocabulary Scales. Oxford, United Kingdom: Oxford Psychologists Press, Ltd.

  • 14

    Hadidjaja P, Bonang E, Suyardi MA, Abidin SA, Ismid IS, Margono SS, 1998. The effect of intervention methods on nutritional status and cognitive function of primary school children infected with Ascaris lumbricoides. Am J Trop Med Hyg 59 :791–795.

    • Search Google Scholar
    • Export Citation
  • 15

    Gardner JM, Grantham-McGregor S, 1996. Baddeley A. Trichuris trichiura infection and cognitive function in Jamaican school children. Ann Trop Med Parasitol 90 :55–63.

    • Search Google Scholar
    • Export Citation
  • 16

    Checkley W, Epstein LD, Gilman RH, Black RE, Cabrera L, Sterling CR, 1998. Effects of Cryptosporidium parvum infection in Peruvian children: growth faltering and subsequent catch-up growth. Am J Epidemiol 148 :497–506.

    • Search Google Scholar
    • Export Citation
  • 17

    Guerrant RL, Schorling JB, McAuliffe JF, de Souza MA, 1992. Diarrhea as a cause and an effect of malnutrition: diarrhea prevents catch-up growth and malnutrition increases diarrhea frequency and duration. Am J Trop Med Hyg 47 :28–35.

    • Search Google Scholar
    • Export Citation

Author Notes

Reprint requests: William A. Petri Jr., Division of Infectious Diseases and International Health, University of Virginia Health System, Room 2115 MR4 Building, P.O. Box 801340, Charlottesville, VA 22908-1340, E-mail: wap3g@virginia.edu.
Save