To understand the impact of EE on macronutrient digestion and assimilation, knowledge of the site, extent and timing of EE is necessary for the following reasons; (1) there is a proximal to distal intestinal gradient of brush-border membrane (BBM) enzymes and nutrient transporters [11], (2) intraluminal proteolytic degradation of pancreatic enzymes can occur during duodenum-ileal transit (lipolytic activity is the most susceptible to inactivation) [12], (3) the influence of ontogeny of pancreatic and BBM enzymes with age, and the influence of endogenous hormones and exogenous (diet, environmental stressors) factors [13, 14] and (4) secondary or functional exocrine pancreatic insufficiency (EPI) can occur due to impaired entero-endocrine pancreatic signaling from the proximal intestine [15]. These mechanisms underscore the spatial and temporal relevance of the highly organized intestinal delivery and digestion of macronutrients in the intestine.

There is an established duodenal-ileal gradient for absorption of glucose, AA, and peptides; AA absorption is greater in the distal small intestine than proximal, which is reverse in case of di- or tripeptides [13, 16]. Sucrase-isomaltase and β-glycosidase activities are higher in proximal jejunum, while glucoamylase is high in proximal ileum. Peptide and long-chain fatty acid transporters show predominance in the duodenum and jejunum [11]. Lipase and proteases lose ~35% of their activity during the duodenum-jejunal transit, with further loss (>40%) in transit to the ileum, as observed in response to infused macronutrients [12]. Lipolytic activity is the most susceptible to inactivation compared with proteolytic and amylase activity, and loss of elastase is minimal. The survival of the pancreatic enzymes depends on the interaction between delivery of nutrients, and the location/rate of absorption of nutrients and bile acids [12]. The BBM disaccharidase lactase-phlorizin hydrolase (LPH) activity, which is present along the entire length of the small intestine at birth shows a gradual decline with cessation of breastfeeding in many but not all ethnic groups, whereas sucrose-isomaltose (SI) activity and fructose absorption increases with introduction of complementary foods [13]. Pancreatic amylase activity that is undetectable at 1 month of life, increases subsequently with the introduction of complementary foods reaching adult levels by 2 years of age [14]. It is noted that the intestinal response to a diet depends on the age of the child, the duration, and any previous exposure [13]. Overall, studies set to determine digestive and absorptive capacity in early life (<2 years of age), should ensure that there are age-matched controls and consider evaluation of pre-experimental dietary patterns.

However, there is little information on the extent or intensity of damage of the small intestine in EE, as multiple biopsy sampling across the entire length of intestine is not feasible nor ethical in children [3]. In Zambian adults at least, EE does not affect the ileum (P Kelly, unpublished observations). Thus far, studies have only been able to characterize the degree of gross and histopathological lesions at selected sites of the intestine, mainly the duodenum and jejunum, showing partially or fully atrophied villi with loss of secretory cell lineages or a higher density of intraepithelial infiltrates [3, 4, 17, 18]. Children with severe acute malnutrition (SAM), are known to suffer from a severe form of EE, showing marked duodenal villous atrophy [19]. In addition, recent evidence from rhesus monkeys (EE model), where disrupted colonic barrier explained growth faltering, suggests the possibility of the colon being affected in children with EE [20]. This opens up new questions (Box 1) requiring further investigation into the extent of EE, considering the potential contribution of large intestinal dysfunction to the body’s nutrient economy, through energy harvesting and adaptation [21, 22]. With the advent of new technologies (such as wireless capsule endoscopy) these questions can be investigated in children with EE [1].

The present section attempts to enlist mechanisms (Fig. 1, theoretical possibility or evidence) that can cause macronutrient malabsorption, in EE, followed by available experimental evidence in children (if not, in adults), and discusses gaps in research with a note on how to investigate them.

Fig. 1: The figure illustrates potential mechanisms that could impair digestion or absorption of macronutrients in environmental enteropathy.

Starting from the gastric cavity, hypochlorhydria reduces pepsin activity (may impair digestion) and leads to dysbiosis in the small intestine. Dysbiosis results in bile acid dysmetabolism and reduced fat absorption. Dysbiosis also plays a role in mucosal damage that reduces surface area for absorption, and transporter density. Enterokinase levels may be affected in mucosal damage leading to reduced pancreatic enzyme activity (may impair digestion). Reduced enteroendocrine cell density because of mucosal damage, reduces CCK release that in turn decreases pancreatic enzyme release, which impairs macronutrient digestion. Elevated levels of cortisol or E-coli toxin LPS could also reduce peptide, glucose and AA transport across the brush border membrane. The downward arrow indicates reduction, upward arrow indicates increase in levels. AA amino acids, CCK cholecystokinin, EE enteroendocrine, LPS lipopolysaccharide. Figure created with BioRender.com.

Box 1

Does intestinal damage differ along the small and large intestine?

Do the functional consequences depend on the involvement of the colon?

What is the contribution of nutrient competition for the intestinal microbiota or parasites, to macronutrient malabsorption?

Mechanisms common for all macronutrients

The human intestinal barrier function is dependent on healthy structure and function of the epithelium, mucus layer, secreted antimicrobial factors, and microbiota [23]. Small intestinal bacterial overgrowth (SIBO), an extreme form of dysbiosis, is a well-known entity in EE [4, 24]. Dysbiosis can trigger dysregulated epithelial apoptosis, which could indirectly disrupt the “pore”, “leak” or “unrestricted” paracellular pathways (global barrier loss) [23]. A vicious cycle of hyperimmune response and intestinal inflammation can ensue, and result in perpetuated mucosal damage [23]. This mucosal damage or villous atrophy, as noted in EE, implies a reduced surface area/transporters for nutrient absorption, impaired enterokinase activity (enzyme converting zymogens to their active form) and entero-endocrine-pancreatic signaling leading to EPI and postcibal asynchrony [15, 25]. Overall, these mechanisms could differentially impair digestion and absorption of the macronutrients. Only if EPI is severe enough (trypsin and lipase <10% of the normal output) will it have a significant impact on absorption of protein and fat, reflecting the high functional reserve capacity of the pancreas [26]. Carbohydrate (CHO) digestion is the least affected, as unlike protein or fats it is well compensated by salivary and gastric amylase, with further breakdown (fermentation) by colonic bacteria [15].

Evidence for EPI

There is evidence of EPI being associated with duodenal enteropathy (of varying etiology), in children [27]. Children showing mild to severe partial villous atrophy or flat mucosa or lymphocytic infiltration of lamina propria on duodenal biopsies had pancreatic insufficiency of varying degree, which was determined using fecal elastase-1 assay [27]. Furthermore, apparently normal Senegalese children (mean age ~1 year) at high risk for EE, were reported to have sub-optimal pancreatic exocrine output, with lower amylase (35% of normal), lipase (6.8% of normal), trypsin and chymotrypsin (~40% of normal) when compared to age-and sex-matched (aged ~1 year) French children, which was then termed “silent pancreatic insufficiency” [28]. The possibility of this relatively low enzyme output causing malabsorption is more likely for fat (with <10% lipase output) than for protein or carbohydrates [15]. A study in adults with severe pancreatic insufficiency (<5% of normal enzyme output) reported a higher (7 times, as energy) ileal cumulative nutrient delivery, that is ~40% of the administered easily digestible low-calorie meal was malabsorbed [29]. In addition to this, there was accelerated gastric and small intestinal transit (2 times) and premature transition from fed to inter-digestive motility pattern compared to normal adults [29].

From these findings, EPI seems to be one of the potential mechanisms, depending on severity, by which EE could cause malabsorption of macronutrients. This suggests the possibility of pancreatic enzyme replacement therapy (PERT) to support digestion, in children with moderate to severe EE, and has been tried before. A reduction in mortality (19% versus 38%) but no difference in weight gain was noted on administration of PERT (containing lipase, amylase and protease), at a dose of 3000U of lipase/kg body weight for 28 days, in children (mean ± SD age of 20 ± 12 months) with SAM, in comparison to those who did not receive therapy [30]. In children (age range 6–30 months) with coeliac disease (similar to EE) with sub-normal pancreatic enzyme output as observed in duodenal aspirates, PERT (first 30 days) has shown a significant percentage increase (9 versus 5%) in weight-for-age, in comparison to a control group on placebo [31].

Mechanisms specific for each nutrient has been described under each macronutrient, below. In EE, these mechanisms may operate individually or synchronously, and an additive effect may substantially reduce nutrient assimilation, which may be buffered by compensatory adaptive mechanisms by the host, mainly in the gut.

Carbohydrates Mechanism with supporting evidence

Apart from the general mechanisms described above, a direct inhibitory effect of the E-coli toxin lipopolysaccharide (LPS) has been observed on D-glucose transport, in jejunal mucosa of rabbits [32]. LPS seems to alter receptor affinity at the BBM and the basolateral membrane Na+, K+-ATPase activity [32]. The smallest dose (range used 3 µg/mL to 3 × 10−5 µg/mL) of LPS used in this experiment that had an inhibitory effect on the transporters was 3 × 10−5 µg/mL. The plasma LPS concentration in children (2–17 months at recruitment) with non-responsive (4–6 months of nutritional supplementation) stunting and EE has been noted to be in the range of 3–6 × 10−2 µg/mL2. At this concentration, there could be intestinal mucosal or serosal inhibition of nutrient transporters in these children, as observed in the rabbit model. A direct evaluation of this mechanism is difficult in-vivo, especially in children, however, breath tests using carbohydrate substrates (Table 1) can be used to test malabsorption in children with EE and associations can be drawn to circulating LPS concentrations.

Evidence for carbohydrate malabsorption

Table 2 summarizes studies in children with features of EE on small intestinal biopsy samples, and who have undergone measurements for macronutrient malabsorption. Studies in children, with suspected EE (lacking biopsy confirmation), in whom digestion or absorption of macronutrients were tested, are described under each nutrient class herein. These studies do not particularly investigate the possible mechanism by which EE causes malabsorption (as discussed for each nutrient), but some attempt has been made to pin down probable pathways.

Table 2 Experimental evidence for macronutrient malabsorption in Brazilian children with biopsy confirmed environmental enteropathy.

The dual sugar assay using lactulose (disaccharide) with mannitol/rhamnose (monosaccharide), to test intestinal permeability and passive absorption respectively, is widely used as a proxy diagnostic for EE [33]. A study examining the breath excretion of 13CO2 following an oral dose (2 g/kg) of stable isotope labeled 13C-sucrose breath test (SBT) in asymptomatic Australian Aboriginal children (n = 18, 95% CI of age 8–16 months, mean length-for-age z-score of −1.9), reported significantly lower (4 versus 6%) cumulative percent dose recovered in breath at 90 min (cPDR90) when compared to healthy non-Aboriginal controls (n = 7, age range 4–60 months, length-for-age z-score not given). A significant inverse correlation was noted between cPDR90 and lactulose rhamnose ratio (LRR) (r = 0.67, 95% CI: 0.42–0.82), in the Aboriginal children (with and without diarrhea), which suggests that impaired intestinal epithelial integrity could lead to sucrose maldigestion [34]. On the contrary, Zambian adults with biopsy confirmed EE who underwent an optimized SBT, showed similar 13CO2 excretion over the experimental duration in comparison to their healthy Scottish counterparts [35]. This could suggest adequate sucrase-isomaltase enzyme (SI) activity even in EE, either compensated by the unaffected intestine or an upregulation of activity as an adaptation to EE or a high sucrose diet. In support, transcriptomic studies suggest that the brush border expression of sucrase-isomaltase enzyme (SI) and sodium-glucose transporter-1 (SGLT-1) were upregulated, in duodenal biopsies of children with EE (from Pakistan) as compared to age-matched healthy North American controls [36]. However, the current protocol of SBT discussed herein lacks sensitivity to detect mild to moderate SI deficiency and may require modification (Table 1) for further use.

An earlier systematic review summarizes studies that have tested carbohydrate malabsorption in children with SAM [37]. The review reports reduced disaccharide and monosaccharide absorption, with a higher prevalence of lactose malabsorption. In the studies reported, a variety of methods to measure carbohydrate malabsorption were used, such as, carbohydrate loading tests, fecal reducing substances or acidic pH, and a few had disaccharidase levels measured in jejunal biopsies. The studies with jejunal biopsy lacked histopathological confirmation of EE, and therefore the link between EE and malabsorption was not established. The disadvantages of these methods are provided in Table 1, and they may not reflect the true deficit in digestion/absorption or compensation by the remaining gut that was not tested in the time duration of the protocol. Nevertheless, the above evidence suggests the possibility that the BBM enzyme lactase is more sensitive/vulnerable in an EE setting and may serve as an early or better marker of malabsorption. Therefore, lactose breath tests can be employed with the suggested modification in Table 1, to evaluate disaccharide malabsorption, in children with EE.

Proteins Mechanisms with supporting evidence

The high prevalence of hypochlorhydria in malnourished children could theoretically lead to low pepsin activity and downstream consequences of SIBO [38, 39]. There is a great deal of redundant proteolytic activity in the small intestine, with the proportional contribution of pepsin to protein digestion of only ~10–20%. This has been experimentally demonstrated in adults who have undergone gastric bypass surgeries, in whom normal dietary milk protein digestion was reported [40]. On the contrary, reduced protein assimilation has been noted on gastric acid suppression therapy (with proton pump inhibitors for 2 days), but the quantitative contribution may be unimportant [39].

In addition to the consequences of mucosal damage other factors may be implicated in poor absorption. Elevated circulatory cortisol levels as observed in undernutrition with infection or inflammation may impair jejunal peptide transport as was experimentally observed in broilers receiving dexamethasone [7, 41]. In this stress induced (dexamethasone) dose-response study (administered at 0.1, 0.5, and 2.5 mg/kg body weight), the broilers had altered jejunal mucosal morphology akin to EE and showed reduction of glycylsarcosine (artificial dipeptide) transport in an everted jejunal sac experiment, for all three doses, suggesting lower peptide (PePT-1) transporter activity [41]. A direct inhibitory effect of the E-coli toxin LPS has also been observed on Na+-dependent AA transport (leucine), similar to its action on D-glucose transport [32]. On the contrary, rats infected by Cryptosporidium parvum [42], as in EE [43], show a compensatory post-translational upregulation of PePT-1 during acute infection, which maintained the ex-vivo glycylsarcosine transepithelial flux across the ileal mucosa in comparison to non-infected rats [42]. In summary, the extent of local (gut) or systemic infection or inflammatory state could cause protein malabsorption by directly or indirectly inhibiting BBM transporters, with the potential for compensatory upregulation. There is potential for tests using stable isotope AA tracers or glycylsarcosine (Table 1), to investigate these mechanisms in EE.

Evidence for protein malabsorption

There are very few studies in asymptomatic children with EE on the digestive or absorptive capacity of protein or AAs. A recent study conducted in children (aged between 18–24 months) from urban slums in South India, who were classified using a lactulose rhamnose ratio (LRR) cut-off, into EED (LRR ≥ 0.068) and no-EED (LRR < 0.068), showed no statistically significant difference between groups for the systemic availability (after digestion and absorption) of AAs from dietary protein. The dietary protein source tested in this study was an intrinsically labeled mung bean and uniformly labeled spirulina protein [44]. There was also no difference in true phenylalanine digestibility or its absorption index between these EED groups [44]. One possible reason for this observed indifference could be the functional adaptation of the gut epithelium, by either upregulation of PePT-1 transporters or proteases [42].

A study from the past, conducted in adults, point towards AA malabsorption in EE [45]. In healthy asymptomatic Indian men with histopathological features of EE on jejunal biopsy, the mean glycine (AA) and glycylglycine (peptide) absorption from the upper jejunum was lower by 31% when compared to age matched English men [45]. The mechanism for this finding was not investigated, but the researchers suggested theoretical possibilities as pointed out in the mechanisms common for all macronutrients section (above). Contrarily, the finding of lower AA absorption could otherwise imply the role of adaptation to higher protein intakes (mostly animal source) in the English men. Overall, there is some evidence in adults but not children, to indicate AA malabsorption, in EE. Studies using novel approaches using stable isotope AA tracers or glycylsarcosine (Table 1), to determine AA or peptide malabsorption, could be performed after standardization and validation, in children from different geographical settings, with varying degree of EE, preferably confirmed by biopsy.

Fats Mechanism of fat malabsorption with supporting evidence

Dysbiosis is associated with deconjugation (removal of taurine or glycine) of bile salts (BS) to their constituent bile acids (BA), which could decrease the BS levels below the critical micellar concentration for fat absorption, thus causing fat malabsorption [17]. In the past, the triglyceride load test using margarine was used to assess fat malabsorption in children with biopsy confirmed EE (Table 2). As mentioned in Table 2, children with giardiasis or chronic diarrhea showed only half or one third increase in plasma triglyceride concentration when compared to a control group [17, 46]. The probable mechanistic pathway implicated for reduced fat absorption in these children was the presence of a higher rate of deconjugation of BS in their jejunal aspirates. The reported mean concentration of deconjugated and conjugated BA was 20.1 (SD 15.5) µmol/mL and 18.2 (SD 16.5) µmol/mL, respectively, in children with chronic diarrhea [46]. The level of deconjugated BA in the duodenal aspirates of normal adults (aged 18–45 years) is noted to be <1 µmol/mL [47]. Barring the considerable variation in collection of bile, introduced by time from meals, enterohepatic circulation, synthetic capacity of the liver, and transit time of intestinal contents, the concentration of deconjugated BA seems to be high in these children, suggesting the possibility of causing fat malabsorption.

The Study of Environmental Enteropathy and Malnutrition (SEEM) conducted in Pakistan observed a significant positive correlation (rs = 0.32, 95% CI 0.064, 0.543) between percentage plasma glycocholic acid (a primary conjugated BA) and the total EED histopathological score of Pakistani children (n = 55, aged ~24 months) with EED, who were unresponsive to a 2–3 months nutritional intervention [48]. This finding probably indicates sub-clinical cholestasis, as proposed by the study researchers, which means lower concentration of BS in the intestine. Whereas, in another cross-sectional study conducted in Malawian children aged between 12–59 months, with suspected EE (a lactulose mannitol ratio cut-off of ≥0.15), the total median age adjusted serum BA were significantly lower in children with EE (4.51 versus 5.10 mM/L) compared to those without [49]. Additionally, the proportion of BAs conjugated with taurine instead of glycine was modestly but significantly higher in children with EE, in this study [49]. Both these findings indicate altered bile acid metabolism, but do not directly suggest the possibility of intraluminal deconjugation of BS by SIBO. Overall, this mechanism is not well supported by evidence, although it is likely to occur in children with EE, and therefore needs further investigation, using reliable tests to measure fat malabsorption (Table 1) and linking it to intraluminal (duodenal or jejunal aspirate) BS/BA concentration.

A summary, with gaps in research, and future directions

Multiple interlinking pathophysiological pathways leading to sub-optimal availability of macronutrients are implicated/proposed in EE, which could either act in tandem or in synchrony. There is some experimental evidence of lactose, AA and fat malabsorption in support of these proposed mechanisms. The potential for intestinal plasticity with the available reserve capacity may dampen the impact on growth, and lead to catch-up growth in late childhood or adolescence. On the contrary, the time taken for adaptation may be long enough to cause irreversible deficits in domains (immune system, brain) that have critical time windows for development. Future research should focus on understanding the degree of malabsorption of macronutrient, in different populations of children with EE, by adopting and standardizing available protocols. Where facilities permit, colonic biopsies can be conducted to determine if the large intestine is affected. Additionally, functional colonic contribution to the body’s nutrition economy could be explored. Longitudinal studies to establish causal links between EE related malabsorption and growth faltering are necessary, especially in high-risk settings, for early detection and prevention of deficits. Interventions with PERT, pre-digested fat/peptides could be explored in high prevalence areas to establish whether restoration of key nutrients is beneficial or not. In conclusion, the research gaps identified in this review, paves way for meaningful investigations and interventions in EE.