点击显示 收起
1 From the Department of Pediatrics, Washington University School of Medicine, St Louis, MO (LG, MJM, and KF); the Department of Paediatrics, College of Medicine, University of Malawi, Blantyre, Malawi (LG and MJM); the Department of Pediatrics, Children's Nutrition Research Center (RJS and MJM), and the Department of Pathology (CNO), Baylor College of Medicine, Houston, TX; and the Paediatric Research Centre, Tampere University Hospital, Tampere, Finland (PA)
2 Supported by Cooperative Agreement 58-6250-6001 from the US Department of Agriculture Agricultural Research Service. The sponsors of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. 3 Reprints not available. Address correspondence to MJ Manary, Department of Pediatrics, St Louis Children's Hospital, One Children's Place, St Louis, MO 63110. E-mail: manary{at}kids.wustl.edu.
ABSTRACT
Background: Tropical enteropathy is an asymptomatic villous atrophy of the small bowel that is prevalent in the developing world and is associated with altered intestinal function and integrity. The histology of tropical enteropathy resembles that seen in small-bowel bacterial overgrowth.
Objective: This study tested the hypothesis that treatment of 3–5-y-old Malawian children with the probiotic Lactobacillus GG would improve their intestinal function and integrity.
Design: Clinically healthy children (n = 164) were enrolled in a placebo-controlled, randomized, double-blind trial. Intestinal function and integrity were measured by using the site-specific sugar-absorption test before and after 30 d of treatment with Lactobacillus GG or placebo. The primary outcomes were the ratios of urinary lactulose to mannitol (L:M) and of urinary sucrose to lactulose (S:L) excretion.
Results: Of the 161 children who completed the study, 119 (73%) had tropical enteropathy on enrollment (L:M > 0.10). Children receiving Lactobacillus GG did not differ significantly from the placebo group in the excretion (in % of dose administered) of mannitol (
Key Words: Tropical enteropathy • ratio of lactulose to mannitol • Lactobacillus GG • probiotics • children • Malawi
INTRODUCTION
Tropical enteropathy is a subclinical, diffuse, villous atrophy of the small bowel that is associated with T cell inflammatory infiltration present in large numbers of poor and stunted but otherwise healthy children aged <5 y in the developing world (1–3). This condition of the small bowel is associated with compromised intestinal capacity to absorb macronutrients. For example, apparently healthy populations with tropical enteropathy have been found to have an overall 5% decrease in carbohydrate absorption and a 15% decrease in protein absorption (4, 5). This altered absorptive capacity has been implicated in the growth faltering commonly seen in the developing world. The dual-sugar permeability test is a noninvasive measure of gut integrity that correlates with the severity of small-bowel histologic changes (6), severity of diarrhea (7), micronutrient status (8, 9), and growth by weight and length (10). In The Gambia, for instance, tropical enteropathy accounts for 43% of growth faltering (10). However, site-specific sugar-absorption testing, which measures intestinal integrity, has not been done in children with tropical enteropathy.
The etiology of tropical enteropathy is related to an environmental exposure, but the specific environmental factors have not been identified (11). Hygiene and sanitation conditions and practices, such as the cleanliness of the water supply and fecal contamination of food, are said to play an important role in the enteropathy (12, 13). An animal model of tropical enteropathy has been created by enteric colonization of the small intestine (14). One working explanation for tropical enteropathy is that, under conditions of poor hygiene, enteric organisms contaminate food and water and colonize the small bowel, which results in decreased intestinal integrity via an immune-mediated process.
Lactobacillus GG is a safe, widely used probiotic that survives the acidic environment of the stomach and readily colonizes the gastrointestinal tract of children (15, 16). It promotes recovery from diarrheal disease in children (17). Lactobacillus GG displaces coliforms from the small intestines of patients with colonic resections and improves the patients' clinical symptoms (18). Lactobacillus GG has also reduced small-bowel enteric overgrowth in patients with this pathological condition (19). Administration of Lactobacilli improved gut integrity as measured by the dual-sugar permeability test in children with atopic dermatitis, a condition associated with abnormalities in intestinal barrier function (20).
If tropical enteropathy is the result of small-bowel bacterial overgrowth, Lactobacillus GG may be able to displace these bacteria from the small bowel and restore gut integrity. This study tested the hypothesis that the administration of Lactobacillus GG for 30 d in asymptomatic Malawian children at high risk of tropical enteropathy improves intestinal integrity as measured by the site-specific sugar-absorption test.
SUBJECTS AND METHODS
Study population
The study was conducted in June and July of 2004 in the village of Mwenye in the Machinga district of southern Malawi. The habitual diet of these children is maize-based and supplemented with fish. Most villagers are subsistence farmers, but they have difficulty in growing crops in the sandy soil surrounding neighboring Lake Chilwa. About half of the villagers have access to bore holes, a source of clean water. None of the villagers have a water source within their dwelling or on their premises. Food is prepared outside over open wood fires. Almost everyone in this area lives in mud huts with thatched roofs. Health care in these villages is mainly provided by traditional healers and health surveillance assistants; the latter are government employees with 1 y of health care training in first aid and simple health maintenance. The nearest health center is 40 km from Mwenye.
All children residing in Mwenye who were aged 36–60 mo and without evidence of severe acute malnutrition or severe chronic illness were eligible for the study. Caretakers were able to end the children's participation at any point in person or by not bringing the children for their scheduled follow-up visits.
Written and oral informed consent was obtained from all of the children's caretakers before enrollment. The study was approved by the College of Medicine Research Ethics Committee, University of Malawi, and the Human Studies Committee of Washington University School of Medicine in St Louis, MO.
Study design
The experimental design was a randomized, double-blind, placebo-controlled trial. On enrollment, the caretaker of the participating child was questioned about demographic and sanitation data. The sanitation questionnaire included 15 questions that elicited information on the family's type of water source, where it was located, how much water was used in the home daily, the child's and mother's practices with respect to passing of feces, and the child's and mother's practices with respect to handwashing. Each child was checked for edema, and his or her height, weight, and midupper arm circumference were measured. Then each child completed the site-specific sugar-absorption test. On completion of the test, the children were given a uniquely numbered bottle with 60 powder-filled capsules containing either Lactobacillus GG or placebo. The bottles were prepared prospectively and independent of the field investigators with the use of a random number generator to establish whether the bottle would contain probiotic or placebo. The code indicating which substance each child received was kept in St Louis and not made available to the investigators until the clinical and laboratory analyses were completed and the data were entered into a locked database. All participants were enrolled on 2 consecutive days in June 2004.
Caretakers were instructed to administer the contents of a capsule twice a day by opening a capsule, sprinkling the contents on a small amount of maize porridge at ambient temperature, and then feeding the child the mixture with a spoon. Maize porridge is a staple food consumed every day. Children assigned to the intervention group received 2 capsules/d containing 510 Lactobacillus GG bacilli (Valio Ltd, Helsinki, Finland). The Lactobacillus GG and the placebo were identical in look and taste. The Lactobacillus GG was tasteless and did not affect the characteristics of the food when added to porridge.
During the 30-d study period, each child and his or her caretaker returned twice a week for a follow-up visit. During each visit, a pill count for each bottle was conducted, and the caretaker was questioned about whether the child had cough, fever, diarrhea, vomiting, or any other problems. After 30 d, the children participated in a second site-specific sugar-absorption test and were weighed and measured for midupper arm circumference, and the study concluded.
The primary outcomes for the study were the differences in the fractions of mannitol, lactulose, and sucrose excreted and in the urinary L:M and S:L, as determined by the site-specific sugar-absorption test during the study period. Secondary outcomes were the incidence of fever, cough, and diarrhea during the study period; weight gain during the study period; and the association of sanitation practices with L:M and S:L.
Site-specific sugar-absorption test
Children were asked to take no food after 2200 on the day before the site-specific sugar-absorption test. Each child was met by the research team at a community meeting point in the village at 0600.
Each child first voided into a clean plastic container. This urine was discarded. Each child then drank 100 mL of a solution containing 1 g mannitol, 5 g lactulose, and 10 g sucrose (21). Mannitol, which is a monosaccharide, and lactulose and sucrose, which are disaccharides, are passively absorbed (21). They are not metabolized in the body but, rather, are excreted in the urine. The ratio of the sugars excreted in the urine is related to the permeability characteristics of the gastrointestinal tract. An increase in the urinary S:L compared with that administered reflects the degree of gastric damage (22). Similarly, an increase in the urinary L:M reflects the extent of intestinal damage (23).
Caretakers and children remained at the community meeting point for the next 3 h, during which all of the child's urine was collected in one sterile cup with 10 mg merthiolate added to prevent bacterial degradation of the sugars. The time of ingestion of the sugar solution was marked on the cup to record the record the duration of the urine collection. After 2.0 and 2.5 h, to facilitate the collection of urine, the children were given water to drink. After 3 h, the children were encouraged to void, and the urine collection was completed.
On completion of the urine collection, the total urine volume was measured by using a graduated cylinder. A 4-mL aliquot was placed in plastic cryovial. Urine specimens were transported to the pediatric laboratories at the Medical College of Malawi and stored in a –70 °C freezer. On completion of the study, the urine samples were shipped to the pediatrics and pathology laboratories at Baylor College of Medicine for analysis.
Urinary sugars, lactulose, mannitol, and sucrose were measured by using a modification of the method of Catassi et al as our group described previously (23, 24). Briefly, the urine is spun in a 1.5-mL microcentrifuge tube (Abbott Laboratories, Abbott Park, IL) at 3000 rpm for 5 min at room temperature. A 20-µL aliquot of the urine is injected onto an Aminex HPX 87C 300 x 7.8–mm cation-exchange column protected with a De-ashing Guard Cartridge precolumn (Bio-Rad Laboratories, Richmond, CA) and eluted with degassed pure water at a flow rate of 0.6 mL/min at 85 °C. The column effluent is monitored with a differential refractometer (Millipore Corporation, Mediford, MA) with a sensitivity setting of 128, scale factor setting of 25, and internal temperature of 50 °C. The column is calibrated by using sucrose, lactulose, and mannitol as standards. The interassay CV of this method in our laboratory is 5%.
Statistical analysis
The sample size of 150 children was chosen to detect a 10% increase in mannitol excretion and a 10% decrease in L:M with 80% power and 95% specificity; this assumed that the mannitol and lactulose measurements had a distribution similar to that in a previous Malawian study in which 80% of healthy children had tropical enteropathy (25). It was also assumed that 10% of children enrolled would drop out. Reported normal values of L:M from populations of healthy children in the developed world vary between 0.03 and 0.12 (9, 26, 27). On this basis, an L:M > 0.10 was arbitrarily chosen as abnormal and as being indicative of tropical enteropathy. A normal S:L has not been defined in children. Intention-to-treat analysis was used to compare all subjects' results within the group to which they were originally assigned.
The fraction of the dose of mannitol, lactulose, and sucrose that was excreted and the difference in the fractions of sugars excreted before and after the intervention were calculated for each subject and compared between the 2 groups by using Student's t test. Because L:M and S:L were not likely to be normally distributed, these ratios and the changes in these ratios before and after the intervention were compared by using Wilcoxon's rank-sum test. The rates of weight gain were compared by using Student's t test. The incidences of fever, cough, and diarrhea were compared by using Fisher's exact test. The associations between L:M and S:L on enrollment and the demographic, anthropometric, and sanitation characteristics of the subjects were ascertained by using linear regression modeling (enter mode and backward mode). For every comparison, P < 0.05 was considered significant.
RESULTS
One hundred sixty-four children were enrolled in the study, and 161 completed the study (Figure 1). No vomiting or other adverse effects were reported by caretakers during the study. The demographic, anthropometric, and sanitation practice characteristics did not differ significantly between the Lactobacillus GG group and the placebo group (Table 1). Pill counts were accurate on every visit by all subjects except one, in which case the caretaker reported losing the remaining capsules after 20 d of compliance.
FIGURE 1.. Patient flow diagram.
View this table:
TABLE 1. Demographic, anthropometric, and sanitation characteristics of the study participants
The administration of Lactobacillus GG for 30 d had no effect on mannitol, lactulose, or sucrose excretion or on the urinary L:M or S:L (Table 2). One hundred nineteen (73%) of the children had an L:M > 0.10 on admission. There were no differences in rates of growth or the incidence of fever, cough, or diarrhea between the children who received Lactobacillus GG and those who received placebo (Table 3).
View this table:
TABLE 2. Urinary recovery of lactulose, mannitol, and sucrose in healthy 3–5-y-old Malawian children before and after 30-d administration of Lactobacillus GG or placebo1
View this table:
TABLE 3. Clinical outcomes in healthy 3–5-y-old Malawian children after 30-d administration of Lactobacillus GG or placebo
Fifty-nine children had an L:M > 0.2 on enrollment, which is indicative of a moderate or marked compromise in intestinal integrity. In these 59 children, there was no significant difference in the mean decrease (improvement) in L:M or S:L after 30 d between the 26 children who received Lactobacillus GG and the 33 children who received placebo.
Regression modeling of L:M on enrollment with demographic, anthropometric, and sanitation characteristics showed that the use of a pit latrine and greater water usage in the home were associated with lower, more normal L:M (Table 4). Regression modeling of S:L on enrollment with demographic, anthropometric, and sanitation characteristics showed that the children who washed their hands more often or washed their hands with soap had more normal S:L than did those who washed less often or did not use soap (Table 5). Children who had previously had kwashiorkor had higher, more abnormal L:M and S:L than did children who had not had kwashiorkor.
View this table:
TABLE 4. Association of sanitation characteristics and initial ratio of lactulose to mannitol (L:M) excretion1
View this table:
TABLE 5. Association of sanitation characteristics and initial ratio of sucrose to lactulose excretion1
DISCUSSION
This study found no improvement in intestinal integrity in a population of 3–5-y-old rural Malawian children with a high incidence of tropical enteropathy who received the probiotic Lactobacillus GG for 30 d. Care should be used in extrapolating this finding to other populations or other probiotic regimens. The dose of Lactobacillus GG that we used was sufficient in another study to colonize the gastrointestinal tract in malnourished Peruvian children (28), and thus it was likely that this dose was adequate for this population of Malawian children, although that possibility was not documented with stool cultures. The study team did not directly observe the children receiving the probiotic, but accurate pill counts in both groups suggested good compliance with the regimen.
The only direct measure of small-bowel function and integrity used in this study was the L:M. Other laboratory and histologic measurements of gut function and integrity, eg, serum albumin, fecal elastase and small-bowel biopsy, were not made among these healthy children because doing so was neither ethical nor feasible. However, such additional studies would have provided independent measures of the effects of Lactobacillus GG. The L:M has been shown to respond to interventions designed to enhance gut function and integrity in children within 14 d, and thus it is suitable for this assessment of the effect of a 30-d intervention with Lactobacillus GG (24, 29–31). In addition, we chose the L:M because it has been used extensively in The Gambia (2, 10, 27). To assess gut integrity, other investigators have used other monosaccharides (eg, rhamnose), radioactive EDTA, or polyethylene glycol molecules of various sizes as intestinal probes (32). We chose a safe monosaccharide because the administration of radioactive isotopes is unwarranted in healthy children, and polyethylene glycol probes of small sizes have been shown to be absorbed in an inconsistent manner. There is evidence that both mannitol and rhamnose can permeate cell membranes, that their absorption is affected by the presence of other solutes in the intestine, and that small amounts of mannitol and rhamnose are biochemically degraded by dissimilar mechanisms. Despite these limitations, Travis and Menzies (32) stated that, although there are important differences between the absorption, metabolism, and excretion properties of rhamnose and mannitol, there is "little practical difference between any of the disaccharide/monosaccharide permeability tests" (p 478).
There was no trend of improvement among the children who received Lactobacillus GG. Even among the children who had an L:M of > 0.20 on admission, Lactobacillus GG did not confer improvement in L:M. This suggests that a 30-d regimen of Lactobacillus GG was not effective against tropical enteropathy in this population.
The proportion of children in this study with L:M > 0.10 indicates that tropical enteropathy was prevalent in this Malawian population. These Malawian children had an elevated L:M primarily as a result of mannitol excretion that was about half of the average amount in healthy children living in the developed world (26). Lactulose excretion was high in this study population, but only about 25% above the amounts seen in healthy children in the developed world (26). Low mannitol excretion suggests that a compromised villous surface area contributes relatively more to the tropical enteropathy in this population than do permeable paracellular junctions in the mucosa. Our findings are consistent with those from The Gambia (27).
Gastric permeability as measured by sucrose excretion did not differ significantly between the groups. To our knowledge, there are no data on normal sucrose excretion in children as young as those in our study. In 2 studies carried out in slightly older children, the average sucrose excretion ranged from 0.027% to 0.029% (22, 33). The values obtained from the children in our study appear to be higher (Table 2). Whether this finding truly represents greater gastric permeability requires further study. Furthermore, it is recognized that the expression of sugar excretion as a ratio (eg, L:M or S:L) eliminates the effects caused by differences in gastric emptying or transit time (32). In the previous studies, a ratio was not used (22, 33).
The association between a more normal L:M and the use of a pit latrine and high household water usage in this study supports the observation that unsanitary practices are related to the occurrence and persistence of tropical enteropathy. More normal S:L was associated with more frequent handwashing among children and with their use of soap to wash their hands, which suggests that upper intestinal integrity may be preserved by this practice. A prospective trial of sanitary interventions to ascertain their effect on the L:M and S:L in a vulnerable population would be useful to establish whether these practices do indeed contribute to tropical enteropathy and upper gastrointestinal permeability.
This study is the first report of a trial of a probiotic for the treatment of tropical enteropathy. Lactobacillus GG has been modestly effective in other clinical settings in ameliorating enteric small-bowel bacterial overgrowth (18, 19). The lack of effectiveness of the Lactobacillus GG in this setting should prompt closer examination of the hypothesis that small-bowel bacterial overgrowth causes tropical enteropathy. Studies directly measuring the quantity and type of bacteria present in the intestinal tract of residents of high-prevalence areas by using genetic-profiling techniques may yield further insight into the relation (34). In addition, trials with larger doses or different combinations of probiotic bacteria or broad-spectrum antibiotic therapy to eradicate small-bowel bacteria (or both) may be warranted. Finally, probiotics given to vulnerable children before the onset of tropical enteropathy, before 1 y of age, might prevent the development of the condition.
ACKNOWLEDGMENTS
We thank Petri Raivio for his extraordinary performance as the clinical study manager and Megan and Micah Manary for their field assistance.
LG planned and conducted the study, analyzed the data, and wrote the first draft of the manuscript. KF planned and conducted the study and analyzed the data. RJS and CNO designed the study, performed the laboratory analyses, analyzed the data, and edited the manuscript. PA designed the study and reviewed the manuscript. MJM designed and conducted the study, analyzed the data, and edited the manuscript. None of the authors had any conflicts of interest.
REFERENCES