Acute diarrhea (AD) is usually self-limiting, but in some individuals diarrhea persists for 14 days or more and is defined as persistent diarrhea (PD) (1). In developing countries, 30% to 50% of deaths from diarrhea are due to PD (2). Although the cause of PD is not well understood, several risk factors have been identified and include malnutrition (3), decreased breast-feeding (3,4), infection with enteroaggregative Escherichia coli (EAggEC) (5), concomitant signs of chest infection (3), and decreased cell-mediated immunity (6–8). More recently, it has been shown that in children with AD in whom PD subsequently developed there were increased numbers of peripheral blood CD8+ T cells and a lower CD4+ T-cell response to stimulants than in children who did not have PD (9). With specific diarrheal pathogens, such as rotavirus (RV), we recently found that plasma levels of interferon-γ (IFNγ) but not tumor necrosis factor-α (TNFα) or interleukin (IL)-10, at 6 to 8 days of diarrhea were higher in children who subsequently had PD than in children who recovered without PD (10).
The role of these risk factors is not clear, but they may be interlinked. It has been suggested that PD may not be related to specific biologic factors only but to a series of factors that are biologic, social, and environmental (11). However, because most analyses for risk factors for PD have been conducted in children with similar social and environmental backgrounds, biologic factors may play a key role. Therefore, a better understanding of the underlying biologic factors is required for the prevention and management of PD.
In this study, we examined various aspects of the immune response to assess whether they affect the duration of diarrhea in children with acute watery diarrhea. We compared responses in three groups of children: those with a history of diarrhea for 6 to 8 days who recovered within 14 days of onset; those with a duration of diarrhea of 6 to 8 days in whom diarrhea persisted for more than 14 days after onset, and those who were uninfected (control children). In addition, we assessed the role of clinical features and nutritional status of children with diarrhea in the development of PD.
MATERIALS AND METHODS
Children attending the Clinical Research and Service Centre of the International Centre for Diarrheal Disease Research, Bangladesh (ICDDR,B), in Dhaka, were enrolled. The inclusion criteria for the children were both sexes, 7 to 12 months of age, and presence of watery diarrhea for 6 to 8 days. The exclusion criteria were presence of blood with stools and history of measles in the past 6 months. Children were clinically evaluated by medical history, daily physical examination, and laboratory investigations that included determination of total and differential counts of leukocytes and serum electrolyte levels, when indicated. All children were treated with appropriate fluid replacement therapy. Some children received antibiotics in the hospital if they had concurrent infections, such as respiratory tract infection. Lower respiratory tract infection (LRTI) was defined by the presence of crepitations in the lungs on auscultation.
If, in these children, diarrhea improved within 14 days of onset, the children were classified as having AD, but if diarrhea persisted beyond 14 days, the children were classified as having PD. Improvement in diarrhea was defined as a decrease in stool frequency to less than five in 24 hours and/or change of stool consistency from liquid or loose stools to soft stools. Once PD developed, the children entered the standard management protocol for PD in ICDDR,B, which includes rehydration, dietary management, and the use of drugs where indicated, including micronutrient supplementation. These children were then withdrawn from the study.
Children, also between 7 and 12 months of age, and without any apparent infection for 1 month and no history of measles for 6 months, attending the Nutrition Follow-Up Unit of ICDDR,B, were included as uninfected control subjects.
Stools freshly collected from all children on enrollment were examined microscopically, cultured for enteric bacteria (12), and assessed for the presence of RV by an enzyme-linked immunosorbent assay (ELISA), as described elsewhere (13). The presence of enterotoxigenic (ETEC), enteropathogenic (EPEC), and enteroaggregative (EAggEC) E. coli was determined using specific DNA probes, as described previously (14,15). From children with PD, second samples of stool were collected at 15 to 18 days after the onset of diarrhea and examined for enteropathogens including RV and diarrheagenic E. coli as described for the first sample.
Collection and Storage of Samples
Samples of stool and venous blood were collected from all children on enrollment. Four milliliters of blood was collected aseptically in sterile heparinized tubes (Vacutainer; Becton Dickinson, Rutherford, NJ, U.S.A.), 2 mL in sterile ethylenediaminetetraacetic acid (EDTA)-containing tubes (Vacutainer), and 1 mL in sterile glass vials. Serum was separated from blood in glass vials and plasma from blood in EDTA-containing tubes by centrifugation for 10 minutes and was stored in aliquots at −20°C and −70°C respectively, until assayed.
Fresh blood collected in heparinized tubes was separated on Ficoll Hypaque (Pharmacia, Uppsala, Sweden) by centrifugation at 500 g for 25 minutes. Peripheral blood mononuclear cells (PBMs), which formed a band at the interface, were collected, washed, and counted. Plasma was collected from above the band for PBMs and frozen in aliquots at −20°C. Granulocytes, which formed a pellet with red blood cells (RBCs), were isolated by sedimentation through dextran (Sigma, St. Louis, MO, U.S.A.) and washed twice with a mixture of Hank's balanced salt solution (HBSS, Sigma) and 10 mM 3-(N-morpholino) propane-sulphonic acid (MOPS). The remaining RBCs were lysed by hypotonic shock. Granulocytes were subsequently resuspended at 106 per mL of HBSS-MOPS. In all cases the cell suspension contained more than 95% neutrophils, with the remainder of cells mainly comprising basophils and eosinophils.
Determination of Total Immunoglobulin Levels
Total immunoglobulin (Ig)A, IgM, and IgG in plasma that had been stored at −20°C were measured by turbidimetry in a discrete analyzer (Cobas-Bio; Roche, Nutley, NJ, U.S.A.) with monospecific antisera (Dakopatts, Glostrup, Denmark). Results were expressed as milligrams per milliliters of plasma.
Measurement of Cytokine Levels in Plasma
Cytokines (IFNγ and TNFα) were measured in plasma that was stored at −70°C. Levels of TNFα were measured by enzyme immunoassay (OptEIA Human TNFα kit; PharMingen, San Diego, CA, U.S.A.). The minimum detection limit of TNFα was 1.42 pg/mL. Concentrations were calculated by interpolation from a standard curve and expressed as picograms per milliliter of plasma. Levels of IFNγ were measured with an ELISA, as described elsewhere (10), with a detection limit of 110 pg/mL. Concentrations were calculated by interpolation from a standard curve and expressed as picograms per milliliter of plasma.
Assays for Mononuclear Cell Function
The proliferative response of PBMs was assessed on resting cells by measuring spontaneous DNA synthesis and on cells stimulated with mitogens, as described previously (16). Results of spontaneous DNA synthesis were expressed as counts per minute (cpm). For stimulation, the mitogens used were phytohaemagglutinin (PHA, Wellcome Diagnostics, Dartford, UK) at 0.6 μg/mL, concanavalin A (ConA; Sigma) at 5 μg/mL, and pokeweed mitogen (PWM; Sigma) at 2.5 μg/mL. Results were expressed as a stimulation index that was calculated by dividing the mean count per minute of cells with mitogen by the mean count per minute of cells without mitogen.
Delayed-type hypersensitivity (DTH) responses were measured by skin tests (Multitest CMI kit; Pasteur Mérieux, Lyon, France), wherein seven antigens and a glycerin control solution were injected into the skin of the paravertebral area of the back of each child and the induration measured at the end of 48 hours. An induration of 2 mm or more was considered a positive response. The antigens present in the kit include tetanus (550,000 Mérieux U/mL), diphtheria (1,100,000 Mérieux U/mL), Streptococcus (group C; 2,000 Mérieux U/mL), tuberculin (300,000 IU/mL), Candida albicans (2000 Mérieux U/mL), trichophyton mentagrophytes (150 Mérieux U/mL), and Proteus mirabilis (150 Mérieux U/mL). The test was performed only on those children who had been immunized against diphtheria, tetanus, and tuberculosis.
Functional Assays for Neutrophils
The polarization or shape change assay for neutrophils was performed against log dilutions (ranging from 10−5 to 10−9 M) of the pure chemotactic factor N-formylmethionyl-leucyl-phenylalanine (FMLP; Sigma), as described elsewhere (17). The percentage of polarizing cells was assessed by viewing under a light microscope (BH-2, Olympus, Tokyo, Japan) with ×40 objective. Neutrophils with a spherical outline or with a few filamentous projections were counted as round cells. Any cell deviating from this spherical morphology, including cells with a constriction separating a spherical outline of the membrane from a ruffled outline, was scored as a polarized cell. Polarized cells were expressed as a percentage of the total number of granulocytes counted. At least 300 granulocytes were counted.
Assays were also performed to determine neutrophil attachment to unopsonized and opsonized (in fresh autologous plasma) Baker's yeast particles, as described elsewhere (17). Results were expressed as the percentage of neutrophils attached to at least three yeast particles and the number of yeast particles attached per 100 randomly selected neutrophils. At least 300 neutrophils were counted.
Assays to Determine Nutritional Status
Transferrin was measured in plasma by turbidimetry in a discrete analyzer (Cobas-Bio; Roche) using a transferrin detection kit (Boehringer Mannheim, Mannheim, Germany). Results were expressed as milligrams per decaliter and calculated by interpolation from a standard curve using human serum standards ranging from 59 to 426 mg/dL (Boehringer Mannheim). In control assays a serum protein was used that was provided in the kit, and, in addition, samples were arbitrarily spiked with commercially available serum of known transferrin concentration. The coefficient of variation was 4% to 8%.
Albumin was measured in serum by a photometric calorimetric test kit (ALBUMIN liquicolor, Human Diagnostics, Wiesbaden, Germany) and results expressed as grams per liter.
The Kruskal–Wallis test was used for nonparametric data in comparisons among the three groups of children. For parametric data, the three groups of children were compared using the one-way analysis of variance, and when results were significant, the Bonferroni test was applied to determine differences between any two groups. Two groups were compared using the Mann–Whitney test (for nonparametric data) or the t-test (for parametric data). For comparisons between proportions, the χ2 statistic or Fisher's exact test was used. Differences were considered to be significant when P < 0.05. The correlation between the information gathered at admission from patients with diarrhea and the development of PD was examined with logistic regression. All variables that could be correlated with the development of PD (age, weight for age, stool frequency, presence of LRTI on admission, plasma transferrin levels, serum albumin levels, plasma TNFα levels, DTH response to any antigen, and DTH response to tuberculin) were analyzed. For the regression analysis, only data from patients with diarrhea were used (AD and PD). Data analyses were performed by computer (Statistical Package for Social Sciences, ver. 8.0 for Windows, SPSS, Inc., Chicago, IL, U.S.A.; and Epi Info ver. 6, USD, Stone Mountain, GA, , U.S.A.).
The Ethics Review Committee of ICDDR,B approved the study.
Characteristics of Children Enrolled
A total of 136 children 7 to 12 months of age were enrolled. Of these, 13 were uninfected control subjects, 85 had AD, and 38 had PD. Enteric pathogens isolated on enrollment from the stools of children with diarrhea (i.e., at 6–8 days of diarrhea) and from the stools of children in whom PD developed at 15 to 18 days of diarrhea are shown in Table 1. The only organism that was significantly associated with the development of PD was EAggEC (P = 0.029). When compared with the pathogens isolated from the stools on enrollment, of the 38 children with PD, 17 had either no pathogens or the same pathogen isolated at 15 to 18 days of diarrhea.
The clinical characteristics and nutritional status of the children on enrollment are shown in Table 2. Uninfected children were older than children with diarrhea, whether AD or PD (P < 0.05 for both), but children with AD and those in whom PD subsequently developed were similar in age. The number of children with signs of LRTI on enrollment was similar in children with AD and those with PD. However, when the number of children who had LRTI at any time during the study period was compared, it was found that more children with PD had LRTI (16/38) than children with AD (19/85;P = 0.031). The extent of dehydration was similar in children with AD or those in whom PD developed (data not shown). There were no differences in the total and differential leukocyte counts in the blood among the three groups of children and in the number of leukocytes and erythrocytes in the stools of the two groups of children with diarrhea (data not shown).
The nutritional status (Table 2), measured by weight for age as a percentage of the National Center for Health Statistics (NCHS) median; serum albumin levels; and plasma transferrin levels were similar among the three groups of children.
Table 3 shows the percentage of polarized neutrophils with and without FMLP. The percentages of neutrophils that polarized in response to 10−8, 10−7, and 10−5 M FMLP were significantly higher in children with diarrhea, whether AD or PD, compared with uninfected children. Children with AD and those with PD had similar responses.
The percentage of neutrophils attached to yeast particles and the number of yeast particles attached per 100 neutrophils were similar in the three groups of children (data not shown).
Total Immunoglobulin Levels in Plasma
There were no differences in the levels of total IgA (uninfected children: median = 0.7 mg/mL, 25th–75th quartiles = 0–1.1 mg/mL; AD: median = 0.7 mg/mL, 25th–75th quartiles = 0–1.0 mg/mL; PD: median = 0.7 mg/mL, 25th–75th quartiles = 0–1.0 mg/mL), IgG (uninfected children: median = 10.6 mg/mL, 25th–75th quartiles = 9.1–12.4 mg/mL; AD: median = 8.2 mg/mL, 25th–75th quartiles = 6.1–10.8 mg/mL; PD: median = 7.8 mg/mL, 25th–75th quartiles = 5.7–10.6 mg/mL), and IgM (uninfected children: median = 1.4 mg/mL, 25th–75th quartiles = 1.1–1.9 mg/mL; AD: median = 1.8 mg/mL, 25th–75th quartiles = 1.3–2.3 mg/mL; PD: median =1.5 mg/mL, 25th–75th quartiles = 1.1–2.0 mg/mL) in the plasma of the three groups of children.
Cytokine Levels in Plasma
The levels of TNFα and IFNγ in the plasma are shown in Figure 1, A and B, respectively. There were no differences in the levels of both cytokines among the three groups of children. However, for TNFα, there were three outliers in the PD group (Fig. 1A) that may have affected the analysis and because of which comparisons were repeated without the three outliers. Comparisons without outliers showed significant differences among the three groups of children (P = 0.013). and levels were higher in children with AD when compared with those in whom PD subsequently developed (median = 0 pg/mL, 25th–75th quartiles = 0–2.9 pg/mL;P = 0.017) or uninfected children (P = 0.036). Uninfected children and those in whom PD subsequently developed had similar levels of TNFα in the plasma.
Mononuclear Cell Responses
Spontaneous proliferation of PBMs and proliferation of PBMs in response to mitogens including PHA, PWM, and ConA were compared among the three groups of children. There was a significant difference in the spontaneous proliferation of PBMs among the three groups of children (P = 0.001). Lower proliferation was observed in uninfected children (median = 849 cpm, 25th–75th quartiles = 499–1179 cpm) than in those with AD (median = 2043 cpm, 25th–75th quartiles = 1454–4194 cpm;P < 0.001) or those with PD (median = 1941 cpm, 25th–75th quartiles = 957 –3150 cpm;P = 0.011). Spontaneous proliferation of PBMs was similar in children with AD and in those with PD. There were no differences in proliferation of PBMs in response to any of the mitogens tested among the three groups of children (data not shown).
Delayed-type hypersensitivity responses were compared by the number of children who had a positive response to at least one antigen and to individual antigens (Table 4). Such comparisons showed a significant difference in the tuberculin response among the three groups of children (P = 0.021) and more children with PD had a negative tuberculin response than did children with AD (P = 0.024).
Of the variables analyzed to assess the relative risk (odds ratio, OR) for the development of PD, only a negative DTH response to tuberculin significantly correlated with the development of PD (OR = 3.8, 95% confidence interval [CI] = 1.4–9.9).
In this study we compared clinical features, nutritional status, and immune responses of uninfected children and children with diarrhea at 6 to 8 days after the onset of diarrhea to determine the risk factors for the development of PD. Of all the parameters examined, a negative DTH response to tuberculin was the only parameter that was found to be a significant risk factor for the development of PD. It has been suggested that PD may not be related to specific biologic factors but to a series of biologic, social, and environmental factors (11). Although the socioeconomic status of the children was not assessed in this study, the surveillance data from the CRSC of ICDDR,B (18) shows that patients attending the CRSC are usually from urban areas and were of low socioeconomic status (19).
Uninfected children were older than those with diarrhea (whether AD or PD) but children with AD and those with PD were comparable in age. Because it was possible that older age of uninfected children could influence the analyses, we also conducted the analyses between children with AD and those with PD and the comparisons yielded similar results. Clinical characteristics and routine laboratory tests of the children with diarrhea on enrollment could not differentiate children who recovered versus those who would have PD, which confirm previous data (20). In this study, of the enteric organisms isolated, EAggEC was the only organism that was significantly more common in children in whom PD developed and this corroborates previous findings (5). Also, the enteric organisms isolated after 14 days of diarrhea was in most cases different from the initial infecting organism, which is similar to other reports confirming that PD is more often due to sequential infection with different enteric organisms rather than to persistence of the same organism (1,21,22).
Taniguchi et al. (9) found that during AD, children who subsequently had PD had lower plasma transferrin levels than those who recovered, but there was no difference in their weight for age. In contrast, Mahalanabis et al. (3) found that low weight for age was associated with the development of PD in children with AD. In the current study no differences were found among the three groups of children in the nutritional parameters that we examined. The reason for the discrepancies between these studies is not clear, but an important difference between these studies is the age of the children enrolled. In the present study, children between 7 and 12 months were enrolled, whereas in the other studies children from 1 to 36 months (3) or 7 to 60 months (9) were enrolled. In Bangladesh malnutrition is most prevalent in children between 7 and 24 months of age (23), and it is possible that the difference in nutritional status between children with AD and PD in this age group is a minor one that would require a larger sample size for detection.
It has been reported that signs of concomitant chest infections are a risk factor for the development of PD (3). In this study there was no difference in the number of children with LRTI at the time of enrollment (at 6–8 days after the onset of diarrhea) in children with AD and those in whom PD developed. Although more children in whom PD developed also had LRTI develop during their hospital stay, it is not possible to exclude that similar numbers of children with AD would also have LRTI had they been observed for a similar duration. However, it is known that systemic infections in children with PD are common (24) and are a risk factor for death caused by PD (25). The association between PD and other infections is not clear.
Studies performed prospectively on healthy children from the community (6) and children with diarrhea from the hospital (9,10) have so far shown that altered immune responses precede the development of PD. In this study we examined neutrophil function, mononuclear cell function, total immunoglobulin levels in plasma, IFNγ and TNFα levels in plasma, and DTH responses to skin tests. Of all the immune parameters examined a statistically significant difference between children with AD and those in whom PD developed was found in the DTH response to tuberculin, and a negative response was found to be a significant risk factor for the development of PD. The DTH response to tuberculin is related to the nutritional status; presence of intercurrent viral infections, especially measles, varicella, and influenza; and the immunologic status of the children (26). Progressive human immunodeficiency virus (HIV) with declining CD4+ T cell counts is also associated with a negative tuberculin response (27). In the present study, the nutritional status of the children was similar in all three groups. Of the intercurrent infections, all children enrolled had no history of measles in the past 6 months, and none had any apparent signs of chicken pox. Also, on enrollment, equal numbers of children in the AD and PD groups had signs of LRTI. We did not test for HIV in this study because the data so far show that the prevalence of HIV is still low among population groups exhibiting high-risk behavior in Bangladesh (28) and that among 250 children with PD attending the CRSC of ICDDR,B who were tested for HIV, none were positive (29).
It is likely that some immunologic parameter is altered in children who progress to PD. The tuberculin skin test is a type IV hypersensitivity response in which there is an influx of neutrophils at the initial phase followed by T cells and monocytes (30). Of the T cells, CD4+ T cells predominate. We did not find any alterations in neutrophil polarization, which is an early event in neutrophil activation, or in neutrophil attachment to yeast particles. These findings suggest that altered neutrophil function is unlikely to be responsible for the decreased response in children who have PD. A more likely explanation is decreased activation of CD4+ T cells that has been reported recently in children with diarrhea in whom PD subsequently developed (9). The CD4+ T cells secrete cytokines that can influence the DTH response (30). In this study we did not find significant differences in the plasma levels of TNFα and IFNγ in the three groups of children, and, although, reanalysis of plasma levels of TNFα, after removal of the three outliers, showed significantly lower levels in children who subsequently had PD compared with those with AD, all three children with high TNFα levels had a negative DTH response to tuberculin. This finding, therefore, suggests that the negative DTH response to tuberculin in children who have PD is not related to the secretion of TNFα.
Therefore, in summary, although there is an association between a negative DTH response to tuberculin and the development of PD in children with AD, the mechanisms involved need elucidation.
The authors thank Dr. A. S. G. Faruque and Firdausi Qadri for helpful discussions.
This study was conducted at the International Centre for Diarrhoeal Disease Research, Bangladesh Centre for Health and Population Research with the support of Grant HRN-5986-A-DD-6005-00 from the U. S. Agency for International Development (USAID). ICDDR,B acknowledges with gratitude the commitment of the USAID to the Centre's research efforts.
1. Black RE. Persistent diarrhea in children of developing countries. Pediatr Infect Dis J. 1993; 12:751–61.
2. Synopsis: Persistent Diarrhea Algorithm.
Child Health Research Project; No. 1. Washington, D.C., 1997.
3. Mahalanabis D, Alam AN, Rahman N, et al. Prognostic indicators and risk factors for increased duration of acute diarrhoea and for persistent diarrhoea in children. Int J Epidemiol. 1991; 20:1064–72.
4. Baqui AH, Black RE, Sack RB, et al. Epidemiological and clinical characteristics of acute and persistent diarrhoea in rural Bangladeshi children. Acta Paediatr Suppl. 1992; 381:15–21.
5. Bhan MK, Bhandari N, Sazawal S, et al. Descriptive epidemiology of persistent diarrhoea among young children in rural northern India. Bull World Health Organ. 1989; 67:281–8.
6. Koster FT, Palmer DL, Chakraborty J, et al. Cellular immune competence and diarrheal morbidity in malnourished Bangladeshi children: a prospective field study. Am J Clin Nutr. 1987; 46:115–20.
7. Black RE, Lanata CF, Lazo F. Delayed cutaneous hypersensitivity: epidemiologic factors affecting and usefulness in predicting diarrheal incidence in young Peruvian children. Pediatr Infect Dis J. 1989; 8:210–5.
8. Baqui AH, Black RE, Sack RB, et al. Malnutrition, cell mediated immune deficiency, and diarrhea: a community-based longitudinal study in rural Bangladeshi children. Am J Epidemiol. 1993; 137:355–65.
9. Taniguchi K, Rikimaru T, Yartey JE, et al. Immunological background in children with persistent diarrhea in Ghana. Pediatr Int. 1999; 41:162–7.
10. Azim T, Ahmad SM, Sefat-e-Khuda, Sarker MS, et al. Immune response of children who develop persistent diarrhea following rotavirus infection. Clin Diagn Lab Immunol.
11. Fraser D, Dagan R, Porat N, El-On J, et al. Persistent diarrhea in a cohort of Israeli bedouin infants: role of enteric pathogens and family and environmental factors. J Infect Dis. 1998; 178:1081–8.
12. World Health Organisation. Manual for Laboratory Investigations of Acute Enteric Infections.
Geneva: World Health Organisation; 1987:9–20.
13. Unicomb LE, Bignan F, Rahim Z, et al. A one-year survey of rotavirus strains from three locations in Bangladesh. Arch Virol. 1993; 132:201–8.
14. Albert MJ, Faruque SM, Faruque ASG, et al. A controlled study of Escherichia coli diarrheal infections in Bangladeshi children. J Clin Microbiol. 1995; 33:973–7.
15. Faruque SM, Haider K, Albert MJ, et al. A comparative study of specific gene probes and standard bioassays to identify diarrhoeagenic Escherichia coli in paediatric patients with diarrhoea in Bangladesh. J Med Microbiol. 1992; 36:37–40.
16. Azim T, Sarker MS, Hamadani J, et al. Alterations in lymphocyte phenotype and function in children who develop complications in shigellosis. Clin Diagn Lab Immunol. 1996; 3:191–6.
17. Azim T, Islam LN, Halder RC, et al. Peripheral blood neutrophil responses in children with shigellosis. Clin Diagn Lab Immunol. 1995; 2:616–22.
18. Stoll BJ, Glass RI, Huq MI, et al. Surveillance of patients attending a diarrhoeal disease hospital in Bangladesh. BMJ. 1982; 285:1185–8.
19. Mahalanabis D, Faruque ASG, Albert MJ, et al. An epidemic of cholera due to V. cholerae
O139 in Dhaka, Bangladesh: clinical and epidemiological features. Epidemiol Infect. 1994; 112:463–71.
20. Lanata CF, Black RE, Gilman RH, et al. Epidemiologic, clinical, and laboratory characteristics of acute vs. persistent diarrhea in periurban Lima, Peru. J Pediatr Gastroenterol Nutr. 1991; 12:82–8.
21. Baqui AH, Sack RB, Black RE, et al. Enteropathogens associated with acute and persistent diarrhea in Bangladeshi children <5 years of age. J Infect Dis. 1992; 166:792–6.
22. Lanata CF, Black RE, Maurta D, et al. Etiologic agents in acute vs persistent diarrhea in children under three years of age in peri-urban Lima, Peru. Acta Paediatr Suppl. 1992; 381:32–8.
23. Bangladesh Bureau of Statistics. Child nutrition survey of Bangladesh, 1995–1996, Dhaka, Bangladesh: Statistics Division, Ministry of Planning, Government of Bangladesh.
24. Roy SK, Alam AN, Majid N, et al. Persistent diarrhoea: a preliminary report on clinical features and dietary therapy in Bangladeshi children. J Trop Pediatr. 1989; 35:55–9.
25. Bhutta ZA, Nizami SQ, Thobani S, et al. Risk factors for mortality among hospitalized children with persistent diarrhoea in Pakistan. J Trop Pediatr. 1997; 43:330–6.
26. Starke JR, Jacobs RF, Jereb J. Resurgence of tuberculosis in children. J Pediatr. 1992; 120:839–55.
27. Raszka WV, Moriarty RA, Ottolini MG, et al. Delayed-type hypersensitivity skin testing in human immunodeficiency virus-infected pediatric patients. J Pediatr. 1996; 129:245–50.
28. Azim T, Islam MN, Bogaerts J, et al. Prevalence of HIV and syphilis among high risk groups in Bangladesh. AIDS. 2000; 14:210–11.
29. Fuchs GJ, Ali M, Hossein A, Vermund S. Surveillance of HIV Seropositivity in Bangladeshi Children With Persistent Diarrhoea and Malnutrition.
In: Hawkes S, Qadri F, Rabbani GH, et al., eds. ASCON VI, 6th Annual Scientific Conference, Programme and Abstracts. Dhaka, Bangladesh:. International Centre for Diarrhoeal Disease Research, Bangladesh; 1997:22.
30. Roitt I, Brostoff J, Male D. Immunology.
4th Ed. London: Mosby, 1996.