Secondary Logo

Journal Logo

Original Articles

Nutritional Factors and Helicobacter pylori Infection in Colombian Children

Goodman, Karen J.; Correa, Pelayo*; Tenganá Aux, Heraldo J.; DeLany, James P.; Collazos, Tito§

Author Information
Journal of Pediatric Gastroenterology & Nutrition: November 1997 - Volume 25 - Issue 5 - p 507-515
  • Free

Abstract

Helicobacter pylori infection, an established cause of chronic gastritis and peptide ulcer disease (1,2), is generally acquired during childhood (3). Consistent evidence has implicated H. pylori in the etiology of gastric cancer (4), a neoplasm whose primary risk factors include such dietary components as high salt intake and low consumption of fruits and vegetables (5). The relationship of these dietary risk factors to H. pylori infection has not been studied extensively, nor has the effect of this chronic infection on nutritional status and growth in children received adequate consideration. Given the link between H. pylori infection and low socioeconomic status (3), one would expect the infection to occur more commonly among children with deficient nutritional status, regardless of whether a cause-effect relationship exists. Because poor nutritional status increases susceptibility to infection in general, and infection, in turn, may exacerbate inadequate nutrition, disentangling cause from effect is not an easy task.

The prevalence of H. pylori infection is high among poor children around the world. Results of seroprevalence studies in Africa, Latin America, and parts of Eastern Europe show that 60% to 90% of the population becomes infected during the growing years (6). In the United States, a prevalence of 25% was observed in more than 2,500 children between the ages of 6 and 19 who participated in the National Health and Nutrition Examination Survey from 1988 to 1991 (7); this prevalence was 17% among non-Hispanic whites, 40% among non-Hispanic blacks, and 42% among Mexican Americans. The widespread distribution of this disease in children warrants examination of its relationship to nutritional status.

We conducted a population-based study of H. pylori prevalence and exposures related to infectious disease transmission among children from a rural village in the Colombian Andes (8). This village was selected because of our ongoing studies in this region; these studies were originally motivated by the elevated occurrence of gastric cancer in this population. We reported previously an H. pylori prevalence of 69% among the children of Aldana, increasing from 53% in 2-year-olds to 87% in 9-year-olds (8). Data on nutritional indicators were sought to control for potential confounding of transmission-related exposure effects. Although this study was not designed to measure the effects of the nutritional factors, they have also not been closely examined in published reports. The analysis of their effects is presented here.

MATERIALS AND METHODS

Population

This study was conducted in the Colombian municipality of Aldana, in the Andean zone of the state of Nariño. Natives of Nariño experience elevated rates of stomach cancer and precursor lesions relative to other regions in Colombia (9,10), a country in which the rates of this neoplasm rank among the highest in the world (11,12). Earlier research revealed a prevalence of H. pylori of 93% in a sample of adult volunteers from this region (13). At an altitude of 3010 m, Aldana has an average temperature of 10 °C. In 1992, the population comprised 844 households with a total of 4447 inhabitants, 35% of whom were less than 15 years old. Nineteen percent of the inhabitants lived in close quarters in an urban nucleus. Three years before this study, the Nariño Health Department established in Aldana a comprehensive primary care center with active community outreach programs; the staff of this center collected the data for this study. A more detailed description of the research methods has been published previously (8).

Design

A cross-sectional study was conducted to identify factors associated with H. pylori prevalence. Because the prevalence approaches 100% in the adult population of the region, this investigation focused on the youngest of the children that could feasibly be studied, those 2 through 9 years old. The data were collected in three phases: a household survey of observable exposures related to infectious disease transmission, conducted in all Aldana homes as part of the annual health department census, February to March 1992; an interview of the mothers or primary caretakers of participating children, conducted in the participants' homes, April to July 1992; and screening of the children for H. pylori infection by the 13C-urea breath test, June to September 1992.

Study Population

The child population targeted for inclusion were all current residents of Aldana born between February 1, 1983 and August 1, 1990, spanning the ages of 2 to 9.5 years at the time of H. pylori testing. Eligible children were identified by the annual health department census, which was later determined to have included 96% of the 746 age-eligible children residing in Aldana. Of these 746 children, 685 (92%) completed participation in the study, 29 (4%) migrated from Aldana before participation was completed, and 32 (4%) were not identified by the census and consequently were omitted from the study. One child's breath test was destroyed in processing, leaving a total of 684 children in the analysis.

Definition of H. pylori Status

H. pylori status was determined by the 13C-urea breath test, initially described by Graham et al. (14) and subsequently refined (15). The breath test has become the gold standard of noninvasive diagnostic methods of detecting H. pylori infection (16), its accuracy approximating that of individual biopsy-based techniques. Reported sensitivities and specificities range from 94% to 100% and 93% to 100%, respectively, in persons not recently treated with antibiotics (17-20). The protocol was as follows: A baseline breath sample was collected; consumption of bread and butter to delay gastric emptying was followed by a 2.5-mg/kg oral dose of 13C-urea diluted in a glucose solution; two additional breath samples were taken 20 and 40 minutes later. Gas-isotope-ratio mass spectrometry was used to measure the urea-derived 13CO2 from breath samples. A positive result was defined by an increase in the 13C/12C isotope ratio value of 6/1000 over baseline on either of the subsequent breath samples. Breath tests were rescheduled if the child had taken antibiotics in the preceding week.

Study Factors

This study was designed to investigate exposures of potential relevance to chronic H. pylori infection. Aside from the environmental hygiene-related exposures examined in our previous analysis (8), risk factors for chronic H. pylori infection may include those that affect health status in general and the gastric mucosa in particular. Substances thought to inflict mucosal injury include salt, aspirin, tea, spicy foods, alcohol, and smoking. Ascorbic acid, β-carotene, and such substances as milk that act as buffers may protect the gastric mucosa from injury (21). Respondents were asked to report the types and frequency of fruits and vegetables consumed by the child. The frequency of consuming boiled and unboiled milk was also solicited. Daily intake of vitamin C and β-carotene was determined from the Colombian Institute of Family Welfare Food Composition Table (22) for the reported consumption of milk, fruits, and vegetables, using half of the amount indicated for the average adult portion. Information on intake of coffee, herbal teas, and salt was also obtained. Results of a pilot study revealed that the children in this population were not exposed to active smoking, alcohol, or spicy foods, thus these substances were not examined.

Height-for-age and weight-for-height are the two preferred anthropometric indicators of nutritional status, according to a World Health Organization (WHO) Working Group (23). Low height-for-age can indicate stunting (slowing in skeletal growth), resulting from prolonged inadequate nutrition and chronic or repeated infections, as often occurs in the context of poor economic conditions. Low weight-for-height indicates wasting-that is, a deficit in expected tissue and fat mass for a given height, resulting from failure to gain weight or from weight loss related to limited food intake, infection, or both. Whereas stunting reflects long-term growth retardation, wasting may reflect acute threats to nutritional status. A third commonly used measure, weight-for-age, is less useful as an indicator of nutritional status, because it is a composite of the two preferred measures and does not distinguish tall, thin children from short, well-proportioned children.

Each child's height and weight were measured to determine height-for-age and weight-for-height percentiles on the growth reference curves developed by the National Center for Health Statistics and endorsed by WHO as an international reference (23). Percentiles were obtained using the nutritional anthropometry program of the EpiInfo, Version 5 software program (24). Height-for-age and weight-for-height were classified in two ways: by dichotomous classification based on a cutoff point of two standard deviations below the international median, the cutoff recommended by the WHO Working Group for population comparisons (23); and by multilevel categorization using narrow percentile increments. The multilevel categorization was employed to observe dose-response relationships between height-for-age-weight-for-height and H. pylori prevalence, given that dichotomization based on a conventional cutoff may mask effects that are revealed on comparing the growth distributions more fully (23,25).

Control Variables

To determine whether the study variables were associated with H. pylori infection independently of hygiene-related factors that indicate exposure to modes of infectious disease transmission, and also independently of other aspects of socioeconomic status or general susceptibility to infection, additional variables were included in multivariate data analysis to control for potential confounding effects. To obtain information regarding general susceptibility to infection, questionnaire items ascertained the child's recent medical history-specifically, the type and duration of all illnesses within the preceding 6 months, the number of days with diarrhea in the preceding 30 days, and the type and number of parasitic infections during the preceding 12 months (see Table 1). Although H. pylori-positive children were reported to have been sick a greater number of days on average then H. pylori-negative children, our previous multivariate analysis did not reveal a clear independent association between H. pylori infection status and aspects of the medical history with one exception: a history of amebiasis in the preceding 12 months revealed a strong protective effect (possibly related to effects of antibiotic therapy used to treat amebiasis) (8); therefore, we controlled for this variable in the current analysis. We also obtained information on the intake of aspirin and medications taken for stomach complaints, but medication history was not associated with H. pylori status in our previous analysis. Our original questionnaire did not include information on breast-feeding, therefore, we were unable to examine its effect.

Because of the potentially confounding effects of socioeconomic factors, attempts were made to identify distinguishing socioeconomic variables in this relatively homogeneous setting. Our previous analysis revealed links between decreased occurrence of H. pylori infection and the following socioeconomic indicators: an agricultural job as the principal household occupation (the occupation of highest status among adult household members); gainful employment of the child's mother; and household ownership of rabbits (raised as a source of income) (8). Therefore, we controlled for these variables in the current analysis.

The transmission-related exposures that appeared to be clearly linked to H. pylori status in our previous analysis included number of children and adults in the home, frequency of sharing drinking cups, mother's hand-washing habits, presence of a latrine or toilet in the home and its location relative to the nearest hand-washing facility, drinking water source, frequency of consuming raw vegetables, frequency of swimming in rivers and swimming pools, and contact with sheep (8). These transmission-related exposures were treated as control variables in the current analysis. Given the near universal prevalence of infection among Aldana adults (13) and our previous finding that nearly 90% of Aldana's inhabitants become infected by the age of 9, the infection status of older family members is not a determining factor in this population. The infection status of age-eligible siblings was a strong predictor, however; and its influence was controlled in multivariate regression modeling by the use of a measure of the residual effect of household as described below.

Analysis

Given the cross-sectional study design, the outcome of interest is the prevalence of H. pylori infection, that is, infection present at the time of the breath test. The prevalence odds ratio (OR) was used to estimate the relative risk; the OR compares the odds of having H. pylori infection in an exposed group to the odds of infection in a reference group.

Prevalence odds ratios (OR) were estimated initially by stratified analysis and later by maximum likelihood logistic regression, to adjust for the selected control variables. Product terms were added to the models to detect apparent effect modification by age. To enhance the stability of effect estimates for categorized continuous variables, we chose the category representing the most common exposure level as the referent for the observation of dose-response.

Because the study participants comprise family groups and the outcome is a communicable disease, the data are presumed to violate the standard logistic regression assumption of independent response probabilities across observations. To avoid potential underestimation of standard errors (26), ORs were estimated using the logistic-binomial random effects model for distinguishable data from the EGRET software package (27). This model includes a random effects parameter that is based on the variability of average outcome probabilities across households; this parameter measures a residual household effect on the probability of having H. pylori infection-that is, it measures the effect of the infection status of household members who were included in the study. ORs for the exposures of interest are presented in unadjusted form from univariate standard logistic regression models and in adjusted form from multivariate logistic-binomial random effects models that controlled for the effect of household and potential confounding by the previously mentioned indicators of nutritional status, socioeconomic status, household crowding, and hygiene.

ETHICAL CONSIDERATIONS

The study was approved by the Human Subjects Committee of the Center for the Health Sciences, University of California at Los Angeles. Written, informed consent was obtained from the children and from their parents.

RESULTS

Table 2 presents the distribution of study factors by H. pylori status.

Dietary Factors

As shown in Table 3, the major dietary factors included in this analysis-intake of milk, fruits, and vegetables-along with intake of vitamin C and β-carotene derived from milk, fruits, and vegetables, all displayed an inverse association with H. pylori infection, consistent with a protective effect. Most noteworthy is the strong inverse dose-response effect for the number of fruits and vegetables eaten per day, reflecting an extremely elevated prevalence of infection in children who ate on average less than two daily servings; the OR comparing the extreme categories, less than two daily servings versus six or more, was 30.9 (95% confidence interval (CI) = 5.3-179.7). A similar effect appears for vitamin C intake with high odds of infection corresponding to average daily intakes below 40 mg; children with intakes less than 40 mg per day had 12.5 (CI = 2.5-62.1) times greater odds of infection than children with intakes of 120 mg per day, or more. The prevalence of H. pylori infection appeared elevated in children whose average β-carotene intake was less than 300 IU per day; the OR comparing intakes less than 300 IU per day with 900 IU or more was 3.1 (CI = 1.2-7.9). An inverse dose-response effect, confirmed by categorical analysis, was observed for each 1-cup increase in milk consumption daily, although this estimate was imprecise (OR = 0.79; CI = 0.53-1.16). No difference in effect was observed for boiled versus unboiled milk.

Children in this population frequently drink herbal teas and coffee, usually taking coffee with milk. Consumption of herbal tea showed a weak protective effect; the OR for any consumption on a regular basis versus none was 0.77 (CI = 0.39-1.53), with no apparent dose-response gradient. Coffee taken with milk manifested a protective effect similar to that of milk overall, which includes milk taken with coffee. Black coffee consumption was associated with moderately elevated infection odds; this association became almost null when adjusted for other factors. The amount of salt used in the household for preparing food was not clearly associated with H. pylori infection; the OR comparing somewhat/very salty food with not very salty food, was 0.87 (CI = 0.51-1.46) with no apparent gradient of effect.

Anthropometric Indicators

The distribution of height-for-age in the children of Aldana reflects the predominant short stature of the Andean population; 29% of the children were more than two standard deviations below the median of the international reference. The median height-for-age percentile in the Aldana children was 6.1. Given this distribution, a cut-off defining deficient height-for-age is not clear. Narrow increments of height-for-age percentiles were examined to observe the dose-response. This analysis revealed that the lower three quintiles, including children below the 10th percentile of the international curve, had similar odds of infection, whereas the upper two quintiles had increasingly reduced odds of infection (Table 4); the OR comparing the upper quintile, spanning the 25th to 99th percentiles, with the lower three quintiles was 0.5 (CI = 0.3-1.0). Only 29 children were more than two standard deviations below the international median of weight-for-height; these children had approximately twice the odds of infection compared with other children.

Age-Specific Effects

Although the size of the study population does not yield sufficient statistical power for precise estimates of age-specific effects, a distinctive pattern emerges in examining the effects separately in younger (2-5 years) and older (6-9 years) children (Table 5). All of the dietary factors reveal stronger effects in older children, to greater or lesser degrees. The inverse dose-response trend for milk consumption is clearly stronger in older children. The OR comparing children consuming 2 to 2.5 daily servings of fruits and vegetables with those consuming 3 or more is 3.5 in older children, but just 1.4 in younger children. Comparing the lowest with the highest category of vitamin C intake, the OR increases from 8.3 in younger children to 9.1 in older children. The difference is most dramatic for β-carotene intake, for which a striking dose-response gradient appears in older children, whereas no clear effect appears in younger children. The OR comparing the lowest level of β-carotene intake with the highest level is 10.2 in older children but just 1.8 in younger children. In addition, increased height-for-age appears clearly linked to decreased infection odds in older children, whereas the relationship is not as striking in younger children.

DISCUSSION

The presented findings constitute a strong suggestion that H. pylori infection is related to nutritional status in children, independently of other aspects of socioeconomic status. Specifically, in this Andean population, H. pylori infection occurred more frequently among children of short stature for their age and among those with decreased consumption of fruits, vegetables, and milk. Furthermore, increased intake of dietary vitamin C and β-carotene were associated with a reduced prevalence of infection. These effects persisted after adjustment for indicators of socioeconomic status, household crowding, and hygiene.

Hints that nutrition plays a role in H. pylori infection have appeared in the literature. Several investigators have reported associations between the infection and growth indicators, including short stature (28) or high body mass index (29) in adults and diminished height and weight gain in children (30-33). Among 407 children in Lima, Peru, those who were H. pylori-positive had lower mean values for anthropometric measures as well as for packed red cell volume and serum protein levels (33). Brazilian investigators, however, reported no significant association between nutritional status and H. pylori infection in 241 children of low socioeconomic status (34), although the criteria used to classify nutritional status were not mentioned.

Several lines of evidence suggest that vitamin C may protect against infection. Ascorbic acid inhibits H. pylori ureas activity and growth in vitro (35,36); in humans H. pylori infection is accompanied by low gastric juice-ascorbic acid levels that return to normal after successful eradication of the infection (37); and results of a study of antibiotic treatment failure showed that complaint patients in whom H. pylori infection did not clear had lower baseline plasma and gastric juice-vitamin C concentrations than patients whose infection cleared (38). Dietary data, however, have been scarce. Results of a New Orleans study of low-income gastrointestinal endoscopy patients revealed that patients with low dietary vitamin C intake were more than twice as likely to be infected with H. pylori than those with high intake, while a similar effect was observed for low intake of fruits (39). Follow-up of 341 epidemiologists for several years revealed a decreased risk of H. pylori seroconversion associated with consumption of two or more glasses of milk per day (40). In contrast, an Italian study of 930 adults found little association between selected dietary or plasma nutrients and H. pylori seropositivity (41); however, given evidence suggesting that most H. pylori-positive adults are likely to have become infected as children, the diet during childhood may be more relevant to prevalent infection in adults.

Spontaneous elimination of H. pylori infection has been observed rarely in follow-up of adults with prevalent infection; however, observation of Peruvian infants suggests that it may be common for children in similar settings to undergo repeated cycles of spontaneous eradication and reinfection (42). If this occurs to an important degree, the general limitations of prevalence data apply to the reported findings-in particular, it is not possible to know whether nutritional factors influence the acquisition of the infection or its chronicity. It is not likely, however, that H. pylori infection led to dietary changes in the study population, given that an equal proportion (11%) of both H. pylori-positive and H. pylori-negative children were reported to have changed their diet as a result of stomach complaints. Similar proportions of both H. pylori-positive (16.2%) and H. pylori-negative (16.7%) children were reported to have experienced one or more gastritis symptoms for 3 months or longer. Furthermore, because the breath test results were not available until after the questionnaires were completed, differential ascertainment of exposure status is unlikely.

The dietary factors examined in this report are difficult to measure accurately and are therefore subject to non-differential misclassification-that is, a similar degree of error would occur in classifying the dietary status of H. pylori-positive and H. pylori-negative children, a bias that would tend to underestimate the magnitude of the association between diet and H. pylori infection. Another potential source of misclassification is concern for social appropriateness on the part of respondents. Differential misclassification is a possibility, because the mother's knowledge of appropriate nutrition may have influenced responses and knowledgeable mothers may be more likely to employ behaviors that protect their children from H. pylori. However, given the large ORs and clear effect gradients reported for fruit and vegetable consumption, as well as vitamin C and β-carotene intake, it is unlikely that these associations are entirely caused by differential misclassification.

Given the multiple correlations among the nutritional indicators and control variables, it is not clear whether the effects observed for nutritional factors reflect biologic mechanisms or residual confounding by inadequately measured determinants of infection. However, the extremely strong effects detected for fruits and vegetables and vitamin C intake, which do not weaken when adjusted for several control variables, provide a solid suggestion of a role in preventing chronic H. pylori infection. It should be noted that the effect of fruit and vegetable consumption is far stronger than that observed in our previous analysis for any of the socioeconomic and transmission-related control variables (8). It is therefore unlikely that this effect can be explained entirely by residual confounding.

Observed associations between the prevalence of H. pylori infection and measures of nutritional status are particularly ambiguous in temporality. If these associations do not result from residual confounding by hygiene-related exposures, there are three possible explanations: nutritionally compromised persons are more susceptible to initial H. pylori colonization; they are more susceptible to chronic infection; and/or the infection influences nutritional status. The third possibility does not appear to explain the observed link between H. pylori and diet, given that children in this population did not appear to change their diets differentially by infection status. The observation of stronger dietary effects in older children is compatible with a role of these dietary factors in protecting against chronic infection, although prospective observations are required to distinguish between factors that influence acquisition and persistence of infection. The finding of an age-related change in the apparent effect of height-for-age is particularly difficult to interpret; this observation could reflect prolonged diminished growth resulting from chronic infection initiated at an earlier age or a decreased susceptibility to chronic infection in healthier children, given that height-for-age reflects overall health status.

Despite the limitations, the findings of this population-based study suggest that nutritional factors may be important in determining susceptibility to chronic H. pylori infection. Carefully designed dietary studies should be conducted to determine whether these associations hold in other populations. The potential importance of diet should be considered, not only in investigating the means of controlling this infection, but also in studies designed to identify the role of H. pylori infection in associated disease pathways. In particular, nutrition must be considered in studies of H. pylori infection and gastric carcinoma, given that evidence suggests that diet may play a role in determining both the infection and the cancer.

Acknowledgment: Supported by a grant from the National Cancer Institute and by grants from the Department of Epidemiology and the Latin American Center of the University of California, Los Angeles, U.S.A. The authors thank the staff members of the Aldana Health Center.

REFERENCES

1. Marshall BJ. Campylobacter pyloridis and gastritis. J Infect Dis 1986;153:650-7.
2. Korman MG, Tytgat GN. Helicobacter pylori and peptic ulcer. Scand J Gastroenterol 1995;210:92-6.
3. Goodman KJ, Correa P. The transmission of Helicobacter pylori. A critical review of the evidence. Int J Epidemiol 1995;24:875-87.
4. Schistosomes, Liver Flukes and Helicobacter pylori. IARC monographs on the evaluation of carcinogenic risks to humans. Lyon, France: International Agency for Research on Cancer, 1994;61.
5. Nomura A. Stomach cancer. In: Schottenfeld D, Fraumeni JF, eds. Cancer epidemiology and prevention. 2nd ed. New York: Oxford University Press, 1996:707-24.
6. Pounder RE, Ng D. The prevalence of Helicobacter pylori infection in different countries. Aliment Pharmacol Ther 1995;9:33-40.
7. Staat MA, Kruszon-Moran D, McQuillan GM, Kaslow RA. A population-based serologic survey of Helicobacter pylori infection in children and adolescents in the United States. J Infect Dis 1996;174:1120-3.
8. Goodman KJ, Correa P, Tenganá-Aux HJ, et al. Helicobacter pylori infection in the Colombian Andes: A population-based study of transmission pathways. Am J Epidemiol 1996;144:290-9.
9. Correa P, Cuello C, Duque E. Carcinoma and intestinal metaplasia of the stomach in Colombian migrants. J Natl Cancer Inst 1970;44:297-306.
10. Correa P, Haenszel W, Cuello C, et al. Gastric precancerous process in a high risk population: Cohort followup. Cancer Res 1990;50:4737-40.
11. Pan American Health Organization. Health Conditions in the Americas, 1985-1988. PAHO Scientific Publication No. 524, 1990.
12. Whelan SL, Parkin DM, Masuyer E, Smans M, eds. Patterns of cancer incidence in five continents. Lyon, France: International Agency for Research on Cancer, 1990.
13. Correa P, Muñoz N, Cuello C, Fox J, Zavala D, Ruiz B. The role of Campylobacter pylori in gastro-duodenal disease. In: Fenoglio-Preiser C, ed. Progress in surgical pathology. Philadelphia: Field & Wood, 1989:191-210.
14. Graham DY, Klein PD, Evans DJ, et al. Campylobacter pylori detected noninvasively by the 13C-urea breath test. Lancet 1987;1:1174-7.
15. Graham DY, Klein PD. What you should know about the methods, problems, interpretations, and uses of urea breath tests. Am J Gastroenterol 1991;86:1118-21.
16. Feldman RA, Evans SJW. Accuracy of diagnostic methods used for epidemiologic studies of Helicobacter pylori. Aliment Pharmacol Ther 1995;9:21-32.
17. Klein PD, Graham DY. Minimum analysis requirements for the detection of Helicobacter pylori infection by the 13C-urea breath test. Am J Gastroenterol 1993;88:1865-9.
18. Vandenplas Y, Blecker U, Devreker T, et al. Contribution of the 13C-urea breath test to the detection of Helicobacter pylori gastritis in children. Pediatrics 1992;90:608-11.
19. Logan R, Polson R, Misiewicz J, et al. Simplified single sample 13Carbon urea breath test for Helicobacter pylori: Comparison with histology, culture, and ELISA serology. Gut 1991;32:1461-4.
20. Dill S, Payne-James JJ, Misiewicz JJ, et al. Evaluation of the 13C-urea breath test in the detection of H. pylori and in monitoring the effect of tripotassium dicitratobismuthate in non-ulcer dyspepsia. Gut 1990;31:1237-41.
21. Correa P. Chronic gastritis. In: Whitehead R, ed. Gastrointestinal and oesophageal pathology. 2nd ed. Edinburgh: Churchill Livingston, 1995.
22. Instituto Colombiano de Bienestar Familiar. Tabla de composicion de alimentos colombianos. Subdireccion de Nutricion, Production, y Distribucion de Alimentos. Ministerio de Salud. 1988.
23. WHO Working Group. Use and interpretation of anthropometric indicators of nutritional status. Bull World Health Org 1986;67:929-41.
24. Dean AG, Dean JA, Burton AH, Dicker RC. Epi Info, Version 5: a word processing, database, and statistics program for epidemiology on microcomputers. Stone Mountain, Georgia: USD, 1990.
25. Gorstein J, Sullivan K, Yip R, et al. Issues in the assessment of nutritional status using anthropometry. Bull World Health Org 1994;72:273-83.
26. Greenland S. Modeling and variable selection in epidemiologic analysis. Am J Public Health 1989;79:340-9.
27. EGRET. Epidemiological graphics, estimation and testing package. Statistics and Epidemiology Research Corporation and Cytel Software Corporation, 1991.
28. Mendall MA, Molineaux N, Levi J, Strachan D, Northfield TC. Association of H. pylori with diminished adult height (abstract). Gastroenterology 1994;106:A137.
29. The EUROGAST Study Group. Epidemiology of, and risk factors for, Helicobacter pylori infection among 3194 asymptomatic subjects in 17 populations. Gut 1993;34:1672-6.
30. Weaver LT. Helicobacter pylori infection, nutrition and growth of West African infants. Trans R Soc Trop Med Hyg 1995;89:347-50.
31. Patel P, Mendall MA, Khulusi S, Northfield TC, Strachan DP. Helicobacter pylori infection in childhood: Risk factors and effect on growth. Br Med J 1994;309:1119-23.
32. Raymond J, Bergeret M, Benhamou PH, Mensah K, Dupont C. A 2-year study of Helicobacter pylori in children. J Clin Microbiol 1994;32:461-3.
33. Klein PD, Gastrointestinal Physiology Working Group, Graham DY, et al. Water source as a risk factor for Helicobacter pylori infection in Peruvian children. Lancet 1991;337:1503-6.
34. Oliveira AMR, Queiroz DMM, Rocha GA, Mendes EN. Seroprevalence of Helicobacter pylori infection in children of low socioeconomic level in Belo Horizonte, Brazil. Am J Gastroenterol 1994;89:2201-4.
35. Nilius M, Bode G, Lehnhardt G, Malfertheiner P. In vitro inhibition of Helicobacter pylori urease: Biochemical and ultrastructural analysis. Eur J Clin Invest 1991;21:551-7.
36. Goldie J, Jalali S, Van Zanten S, Stowe C, Hunt RH. Ascorbic acid inhibits the growth and urease activity of Campylobacter pylori. Gut 1989;30:A1484.
37. Phull PS, Green CJ, Jacyna MR. A radical view of the stomach: The role of oxygen-derived free radicals and anti-oxidants in gastroduodenal disease. Eur J Gastroenterol Hepatol 1995;7:265-74.
38. Ruiz B, Rood JC, Fontham ETH, et al. Vitamin C concentration in gastric juice before and after anti-Helicobacter pylori treatment. Am J Gastroenterol 1994;89:533-9.
39. Fontham ETH, Ruiz B, Perez A, Hunter F, Correa P. Determinants of Helicobacter pylori infection and chronic gastritis. Am J Gastroenterol 1995;90:1094-104.
40. Parsonnel J, Blaser MJ, Perez-Perez GI, Hargrett-Bean N, Tauxe RV. Symptoms and risk factors of Helicobacter pylori infection in a cohort of epidemiologists. Gastroenterology 1992;102:41-6.
41. Palli D, Decarli A, Cipriani F, et al. Helicobacter pylori antibodies in areas of Italy at varying gastric cancer risk. Cancer Epidemiol Biomarkers Prev 1994;2:37-40.
42. Klein PD, Gilman RH, Leon-Barua R, Diaz F, O'Brian-Smith E, Graham DY. The epidemiology of Helicobacter pylori in Peruvian children between 6 and 30 months of age. Am J Gastroenterol 1994;89:2196-9.
Keywords:

Child; Colombia; Epidemiology; Growth; Helicobacter pylori; Nutrition

© Lippincott-Raven Publishers