Studies of the Small Intestine in Persistent Diarrhea and Malnutrition: The Gambian Experience

During the past decades, as improvements in treating acute diarrheal disease in children in developing countries have affected morbidity and mortality, persistent diarrhea (diarrhea lasting 14 days or longer), has emerged as a more common cause of death than previously recognized. Information from India, Bangladesh, Senegal, and Brazil suggests that about 45% of diarrhea-associated deaths may be caused by nondysenteric, persistent diarrhea (1). Persistent diarrhea seriously affects growth and nutritional status and presents a very difficult management problem in developing countries. Antecedent malnutrition, lack of breast-feeding, and associated systemic infection increase the risk of death during persistent diarrhea (2). The pathogenesis of persistent diarrhea is complex, and clear evidence suggests microbial infection either through prolonged acute intestinal infection or sequential acute infections. But irrespective of the nature of the initiating factor, a vicious cycle probably will ensue, triggered by a number of factors including the nutritional consequences of poverty, poor hygiene with environmental contamination, faulty feeding practices, and early weaning. Figure 1 clearly portrays the complexity of these interactions. This conceptual model was presented at the first meeting of what was to become the Commonwealth Association of Pediatric Gastroenterology and Nutrition (CAPGAN), an organization established and lead by Professor John Walker-Smith, and was to inspire much research in persistent diarrhea (3). This model points to the fact that whatever the relative contribution of infective, nutritional, or even allergic phenomena in the pathogenesis of persistent diarrhea and malnutrition, the final common pathway probably is prolonged injury to the small intestinal mucosa. Understanding the complex interactions that take place at the small intestinal luminal mucosal interface is central to understanding the nature of persistent diarrhea.FIG. 1.: A conceptual model of the pathogenesis of persistent diarrhea and malnutrition (3).To understand more about the small intestinal damage in persistent diarrhea and malnutrition, we undertook a study in Gambian children who had severe protein energy malnutrition. Forty children (aged 6–36 months), who had more than three loose stools each day for more than 2 weeks and whose weight for height was less than 75% of the National Centre for Health Statistics median value, underwent clinical assessment and jejunal mucosal biopsy on admission and again after 4 weeks of intensive treatment in the hospital (4). A wide spectrum of pathologic changes was observed in the small intestinal mucosa, ranging from near “normal” appearances to a “flat” celiac disease–like lesion. Figure 2 shows jejunal biopsy specimens from each of the 40 Gambian children studied, arranged in descending order of surface epithelial volumes (VSE × 106 μm3) with their corresponding crypt epithelial volumes (VCR) displayed below the horizontal zero reference. The majority (29/40) of specimens showed marked hypertrophy of the crypts that invariably contained an elevated population of intraepithelial lymphocytes (IEL) (NV,CR). In 13 of these specimens (Figure 2, extreme left-hand of the display), surface epithelial volumes (VSE) were normal, although villous epithelium contained an increased number of IELs. The absolute number of IELs in the villous epithelium was proportional to the surface epithelial volume such that flat mucosae contained fewer lymphocytes than more normally structured mucosae. Six specimens at the extreme right-hand of the display were “flat” and lay within the range of VSE for untreated celiac mucosae; all demonstrated crypt hypertrophy, except the last biopsy from the sole patient with kwashiorkor. Immunofluorescent staining of IELs demonstrated that all were CD3+, of which more than 80% were CD8+. Therefore, using a computerized image analysis technique to evaluate the jejunal mucosae from the 40 sick Gambian children, a spectrum of change reminiscent of those seen in other human enteropathies associated with gluten sensitivity, giardiasis, and tropical sprue was observed. The mucosae from some of these children showed features of crypt hyperplasia and hypertrophy with infiltration with CD8+ IELs into crypt and villous epithelium, despite the villi being of reasonable size (5). This lesion contrasts with the milder “infiltrative” lesion, in which lymphocytes occupy villous epithelium only, as occurs in dermatitis herpetiformis and after low-dose gluten challenge. Some of the remaining children demonstrated a flat or “destructive” lesion, characterized by complete lack of villi and infiltration of crypt epithelium that expressed D-locus alloantigens in a similar manner to that seen in celiac disease. In all these various conditions, a stereotyped progression of mucosal change in injury is seen with a striking parallel to the intestinal lesions evoked by various models of allograft and graft-versus-host reactions in experimental animals. These are T-lymphocyte–dependent phenomena brought about by lymphokine-activated cells and restricted by major histocompatibility complex class 2 alloantigen expression (6).FIG. 2.: Morphometric data from 40 Gambian children with persistent diarrhea and malnutrition are displayed in descending order of villous (surface) epithelial volumes, VSE (LH axis) together with their corresponding crypt epithelial volumes, VCR. Horizontal lines represent the lower reference range for control villi, and the upper reference ranges for flat (celiac sprue) mucosae and control crypts. The majority of mucosae display crypt hypertrophy, even in the presence of normal villi, that is, “hyperplastic” mucosal lesions. Accompanying this lesion is infiltration of surface and crypt epithelium by small, intraepithelial lymphocytes, values of which for each specimen are illustrated graphically on RH axes as logarithmic transformations (horizontal solid lines represent respective upper reference ranges). As villous volume decreases, surface intraepithelial lymphocytes progressively decrease, whereas those in crypts tend to be sustained at high levels. The final specimen in this series is the atrophic mucosal in kwashiorkor. (Reproduced from reference 4).These findings are consistent with a cell-mediated intestinal reaction to an environmental antigen, which could be dietary or microbial. A subsequent study further characterized the mucosal inflammatory response in Gambian children (Campbell D, et al., unpublished observations). Morphometric and immunohistochemical analysis was performed for a range of lineage and activation markers, including proinflammatory and regulatory cytokines, and these were related to clinical status and gut permeability using a dual sugar permeability test. All Gambian children studied, regardless of nutritional status, had evidence of chronic cell-mediated enteropathy with crypt hyperplasia, villus stunting, and high numbers of IELs. Numbers of activated T-lymphocytes were increased 20-fold compared with United Kingdom controls and, although small bowel architecture was independent of nutritional status, T-cell numbers increased and B-cell numbers decreased with worsening nutrition, with evidence of decreased activation. Mucosal cytokine production was biased toward a proinflammatory response with a progressive imbalance between proinflammatory and regulatory cytokines as nutritional status worsened toward marasmus (child's weight for height less than 75% of the National Centre for Health Statistics median value). The enteropathy associated with poor nutrition, therefore, seems to predate the onset of marasmus and is essentially an uncontrolled Th1 response. Protein energy malnutrition is associated with a progressive proinflammatory bias of the mucosal cytokine microenvironment in which the known mechanisms of oral tolerance to dietary antigen seem to be disrupted. Previous studies using dual sugar permeability tests confirmed that damage to the small intestinal mucosa plays an important part in the development of persistent diarrhea and malnutrition in The Gambia (7). Another Gambian study extended these findings and aimed to evaluate mucosal morphology and intestinal permeability in Gambian persistent diarrhea and malnutrition to determine to what extent these were reversed by intensive inpatient nutritional rehabilitation (8). This is the only study of serial small intestinal mucosal architecture in children recovering from persistent diarrhea and malnutrition in the tropics, and the results highlight the significance of prolonged injury to the small intestine in children with persistent diarrhea and malnutrition. Twenty children with persistent diarrhea and malnutrition were assessed before inpatient treatment for 4 weeks with antibiotics and high-nutrient–density milk. Intestinal permeability (measured by lactulose:mannitol absorption) and mucosal structure (measured by quantitative, computerized morphometry) were evaluated in 20 children (11 boys and 9 girls, aged 6–31 months) before and after inpatient rehabilitation from persistent diarrhea and malnutrition. Subjects received the ward's standard nutritional rehabilitation diet for malnourished children, combining a high-calorie milk formula with added minerals and a multivitamin supplement. Each child received up to 180 kcal/kg daily as energy supply and 4g/kg protein daily. In addition, all received antibiotics as routine and antihelminthic therapy as indicated. Mean (range) weight-for-height on presentation was 65% (52–77%) of the National Centre for Health Statistics median value, and 1 month after nutritional rehabilitation and treatment of infection, all but two patients had gained weight. In one of these, who presented with kwashiorkor and had the severest mucosal flattening of the series, weight loss represented resolution of edema and the child ultimately recovered completely. Criteria for discharge (cessation of diarrhea for 5 consecutive days and steady weight gain) were met in all subjects within 4 weeks. Lactulose-to-mannitol ratios were high on admission, (0.66 ± 0.36) and, despite some temporary improvement, did not significantly improve (0.49 ± 0.30) after treatment for 1 month. The changes in lactulose-to-mannitol ratio were largely because of an increase in lactulose absorption, suggesting that the small intestinal mucosa becomes more “leaky” as a result of nutritional restoration. Figure 3 displays the results of quantitative morphometric analysis of paired jejunal biopsy specimens from 20 children before and after inpatient treatment of persistent diarrhea. In the majority of cases, morphometric variables between the paired biopsies did not change significantly, although nutritional restitution was associated with a significant increase in the size of the mucosal crypt cell compartment, but not in villous epithelial volumes during the same period. In at least three cases, there was clear evidence of deterioration in mucosal appearance with further loss of villous epithelium. Although no formal correlation was found between measures of intestinal permeability and mucosal morphology, lack of improvement or deterioration in intestinal mucosal architecture is consistent with simultaneous evidence of mucosal leaking. At the end of the inpatient rehabilitation period, fecal α-1-antitrypsin excretion and intestinal permeability increased, as measured by the lactulose-to-mannitol permeability test. Interestingly, the four patients who demonstrated distinct improvement in mucosal appearances initially had the most severe mucosal damage. This suggests that at least in these patients, nutritional rehabilitation was associated with mucosal regeneration. After discharge, progress was again assessed clinically and anthropometrically at 6 and 12 months. Recurrence of diarrhea was common in these children after return to their villages. Nevertheless, continuing improvement in weight for age and mid–upper-arm circumference was observed after 6 and 12 months, and weight for height continued to improve up to 6 months but decreased again by 12 months. Ten children from the original group again underwent jejunal biopsy at 1 year, and all had appearance and measurements of the jejunal mucosa that indicated deterioration in mucosal architecture when compared with the initial biopsy specimens. In conclusion, observation of these patients demonstrated that in treating persistent diarrhea and malnutrition in developing countries, relatively short periods of inpatient rehabilitation, although leading to resolution of diarrheal symptoms and weight gain in the short-term, do not lead to complete recovery. Persistent diarrhea and malnutrition likely will recur when the child is discharged home to an environment characterized by poor sanitation and inadequate water and food. A prospective study must establish the minimum period of supervised feeding required to ensure complete restitution of small intestinal mucosae to normal.FIG. 3.: Morphometric data from 20 Gambian children with persistent diarrhea and malnutrition before and after nutritional rehabilitation are displayed in a similar fashion to Figure 2. Villous (surface) epithelial volumes, VSE (LH axis) together with their corresponding crypt epithelial volumes, VCR are shown. Horizontal lines represent the lower reference range for control villi, and the upper reference ranges for flat (celiac sprue) mucosae and control crypts. Values for absolute numbers of surface and crypt intraepithelial lymphocytes for each specimen are illustrated graphically on the RH axes as logarithmic transformations (horizontal solid lines represent respective upper reference ranges). In the majority of cases, morphometric variables did not change significantly between the two biopsy specimens. At least three patients showed clear evidence of deterioration in villous epithelium. Only four patients had a distinct improvement in mucosal appearance.