The Risk of Necrotizing Enterocolitis Differs Among Preterm Pigs Fed Formulas With Either Lactose or Maltodextrin

Abstract

Objectives: 

When breast milk is unavailable for preterm infants, formulas are needed that won’t increase the risk of necrotizing enterocolitis (NEC). Adding novel ingredients to formula to reduce NEC has not been effective clinically. Instead, we tested the prediction that NEC can be reduced by removing the maltodextrin now included in preterm formulas.

Methods: 

The preterm pig model of spontaneous NEC was used to evaluate growth, health, and intestinal responses to 6 to 7 days of feeding formulas that were identical except for the source of carbohydrate; either 100% lactose or maltodextrin; colostrum was used as the control.

Results: 

Formula with maltodextrin resulted in a 50% incidence of NEC with 30% mortality. The lactose formula and colostrum resulted in a 0% incidence of NEC. Growth was highest for pigs fed the formula with lactose, intermediate with maltodextrin, and minimal when bovine colostrum was fed (P < 0.05). Although the small intestine was larger when colostrum was fed (P < 0.05), because rates of glucose uptake were lower (P < 0.05), total small intestine capacities to transport glucose were similar for healthy pigs in all 3 groups.

Conclusions: 

If lactose-based formulas reduce NEC clinically, the transition of preterm infants to enteral feeding can be accelerated, improving growth and development, and shortening reliance on parenteral nutrition. Although colostrum protects against NEC, chronic feeding does not promote body weight gain after preterm birth. The preterm pig can be used for preclinical studies that evaluate the mechanisms by which carbohydrates and other ingredients influence growth, development, health, and risk of NEC.

What Is Known

• Extremely preterm infants have an increased risk of necrotizing enterocolitis when fed current preterm formulas that include maltodextrin.
• Attempts to improve existing formulas by adding bioactive molecules present in breast milk have not decreased necrotizing enterocolitis.
• Despite incomplete digestion by preterm infants, the lactose present in breast milk does not pose an obvious risk for necrotizing enterocolitis.

What Is New

• Including maltodextrin in formula fed to preterm pigs increases the risk of necrotizing enterocolitis and reduces growth.
• Formula with only lactose did not cause necrotizing enterocolitis and improved growth.
• Colostrum protects against necrotizing enterocolitis, but chronic feeding may not support adequate growth.

Although the percentage of infants that develop necrotizing enterocolitis (NEC) is declining (Vermont Oxford Network, 2017), numbers of NEC cases are increasing because more infants are surviving birth at <28 weeks of gestation. Despite representing <15% of preterm births, this growing population of extremely preterm infants is responsible for approximately 60% of NEC cases ⁽¹⁾. The heightened risk of NEC results in extremely preterm infants receiving minimal or no enteral nutrition for the first days after delivery ⁽²⁾, and particularly when breast milk is not available. A consequence is compromised postnatal growth ⁽³⁾ which contributes to the risk of disability ⁽⁴⁾, with neurodevelopment particularly sensitive ⁽⁵⁾. Furthermore, the continued reliance on parenteral nutrition (PN) increases the risk of septicemia and other complications.

Aggressive enteral nutrition support to prevent extrauterine growth retardation ⁽⁶⁾ will benefit from preterm formulas that won’t increase the risk of NEC. Unfortunately, adding purported protective components of human milk to formulas has not reduced NEC clinically, especially for infants born at <28 weeks. We pursued a different strategy of identifying currently used ingredients that may increase the risk of NEC, with the focus on the maltodextrin (MDX; corn syrup solids) currently added to formulas ⁽⁷⁾ following the recommendations of the American Academy of Pediatrics ⁽⁸⁾ and the European Society of Pediatric Gastroenterology, Hepatology, and Nutrition ⁽⁹⁾. The majority of studies reviewed for these recommendations were conducted before 2000 and included preterm infants >28 weeks of age and at lower risk of NEC. We predicted from previous results ⁽¹⁰⁾ the elimination of MDX from formula will reduce NEC and improve growth. The ethical constraints associated with using preterm infants as test subjects led us to test this prediction using preterm pigs as a relevant and translational large animal model for nutrition and gastrointestinal development ⁽¹¹⁾ and NEC ^(10,12).

MATERIALS AND METHODS

This study used preterm pigs (n = 31; both sexes) harvested at 91% term (day 105 of 115 day gestation) from 2 timed pregnancy, specific pathogen-free sows of the same genetic lineage. Preterm pigs at this stage of gestational development are similar to approximately 32-week preterm infants and do not require mechanical ventilation, which is a confounding factor for NEC risk among preterm infants ⁽¹³⁾. Details for the caesarian section, the 8 to 9 days of postnatal care, parenteral and enteral nutrition support, assessments of growth and health, and measurement of glucose uptake are provided as Supplemental Digital Content (https://links.lww.com/MPG/B71 and https://links.lww.com/MPG/B72). All aspects of the project that involved the care and use of the experimental preterm pigs were approved by the University of Memphis Institutional Animal Care and Use Committee and were based on the protocol for causing spontaneous NEC ⁽¹⁰⁾.

RESULTS

Growth

Birth weights were similar and averaged 988 ± 100, 998 ± 66, and 1021 ± 67 for pigs fed colostrum (COL) and the 2 formulas with either lactose (LAC) or MDX as the sole sources of carbohydrase, respectively. All pigs gained an average of 12% of birth weight during the 48 hours of PN. Although absolute body weights did not differ at scheduled necropsy (COL = 1051 ± 108; LAC = 1188 ± 73; and MDX = 1159 ± 86), rates of growth during the period of enteral nutrition (EN) were higher for LAC compared with MDX pigs, with both higher than COL pigs (Fig. 1).

FIGURE 1: Daily body weight gains (calculated as the percentage change from the previous day's weight) of preterm pigs during the 6 to 7 days of feeding bovine colostrum or formulas with either lactose or maltodextrin as the sources of carbohydrate, but were otherwise identical. Values are means and standard errors. The different letters indicate significant differences were detected for all comparisons (P < 0.05).

Necrotizing Enterocolitis Incidence

All of the COL and LAC pigs survived to scheduled necropsy and none had lesions or evidence of NEC. The MDX pigs had a 50% incidence of NEC that either resulted in humane euthanasia within the first 2 days of feeding (4/14; 28%) or the presence of mild NEC lesions that were detected at necropsy without overt clinical symptoms (3/14; 21%). The 4 pigs that were euthanized early had developed the characteristic bloating, diarrhea, lethargy, and labored breathing, were unresponsive, and had an average NEC score of 2.8 ± 0.3, with the highest value in the proximal colon (4.3 ± 0.5). The NEC observed in the 3 pigs that survived to scheduled necropsy was considered mild based on an average score of 1.6 ± 0.1 and no lesions that exceeded a score of 3. None of the 3 pigs exhibited clinical symptoms that indicated that NEC was present (subclinical).

Clinical Chemistries

Treatment differences were detected for the majority of measured analytes (Table 1), with the greatest differences between the COL pigs and those fed the 2 formulas. Notable were the higher values of COL pigs for sodium, chloride, blood urea nitrogen (BUN), and the serum lipids. There were no differences for comparisons restricted to LAC and healthy MDX pigs, whereas differences (significant and tendencies) were detected between LAC and MDX pigs with NEC. The only significant difference between MDX pigs with and without NEC was the lower serum glucose levels of the pigs with NEC.

TABLE 1: Serum chemistries for preterm pigs fed colostrum or formulas with lactose or maltodextrin

Intestinal Characteristics

Absolute intestinal mass did not differ among groups (Table 2), but when normalized to body weight the small intestines of COL pigs (50.8 g/kg ± 1.7) were proportionally heavier (P < 0.05) compared with values for the LAC and MDX pigs that were considered to be healthy at necropsy (40.6 ± 1.3 and 42.8 ± 2.3, respectively).

TABLE 2: Small intestine and organ weights for preterm pigs after 6 or 7 days of enteral nutrition with colostrum or formulas with lactose or maltodextrin as carbohydrate sources, carbohydrate, but were otherwise identical in composition

Because NEC reduces glucose uptake ⁽¹⁰⁾ only intestines from healthy MDX pigs without NEC lesions were included in the comparison (n = 7). Rates of glucose uptake (nmol · mg⁻¹ · min⁻¹) averaged for the 3 regions of the small intestine for COL pigs tended to be lower (P < 0.10) relative to the LAC and healthy MDX pigs, whether measured at the saturating concentration of 50 mmol/L or at tracer concentration (Fig. 2). Despite lower rates of uptake per unit of intestine for COL pigs, because of their relatively larger smaller intestines, estimates of total intestinal glucose uptake capacities calculated as the product of rates of uptake multiplied by small intestine mass and normalized to body mass (8.1 g · kg⁻¹ · day⁻¹ ± 1.2) were not significantly lower compared to LAC and healthy MDX pigs (9.2 ± 3.7 and 10.7 ± 1.5).

FIGURE 2: Rates of glucose uptake (nmol · mg⁻¹ · min⁻¹) measured at 50 mmol and tracer concentrations and averaged for the entire length of small intestine for pigs fed bovine colostrum or formulas with either lactose or maltodextrin as the sources of carbohydrate, but otherwise identical. Bars represent means and standard errors.

Other Organs

COL pigs had smaller brains and spleens than LAC and healthy MDX pigs (P < 0.05). Organ weights (absolute and normalized to body mass) did not differ among treatments for the liver, kidneys, and pancreas, suggesting these organs grew in proportion with gains in body mass, whereas the smaller COL pigs had relatively larger hearts and lungs compared with the LAC and healthy MTX pigs (P < 0.05) (Table 2).

DISCUSSION

Although the risk of NEC associated with formula can be reduced by gradually increasing the volume fed to preterm infants ⁽¹⁴⁾ and preterm pigs ⁽¹⁵⁾, the risk still exceeds feeding breast milk to infants and COL to pigs. We now report the source of carbohydrate in formula differentially influences the risk of NEC for preterm pigs with the incidence and severity related to the amount of MDX (10; present study).

Maltodextrin Increases Risk of Necrotizing Enterocolitis

The 30% NEC mortality in the present study for the MTX pigs is comparable to previous studies that fed preterm pigs formula with MDX for 48 hours after a comparable period of PN ^(11,12). Even the addition of ingredients considered to be protective to formula with MDX, such as human milk oligosaccharides ⁽¹⁶⁾ and lactoferrin ⁽¹⁷⁾ did not prevent NEC among preterm pigs. However, the total incidence and severity of NEC after 6 or 7 days was lower than what is usually detected after approximately 48 hours of feeding. Preterm pigs can be asymptomatic, despite the presence of lesions, similar to extremely preterm infants that are asymptomatic until NEC becomes severe ⁽¹⁸⁾. It is possible some pigs were asymptomatic, but had developed lesions that were partially or fully resolved after the 6 to 7 days of feeding.

The higher incidence of NEC for preterm pigs fed formulas with MDX, even at 15% ⁽¹⁰⁾ compared with preterm infants fed current preterm formulas may be related to the species differences for developmental trajectories and activities of brush border membrane carbohydrases. Lactase develops relatively later in human fetuses ⁽¹⁹⁾ compared with pigs ⁽²⁰⁾, whereas human fetuses express sucrase-isomaltase earlier than pigs. As a consequence, preterm human infants have greater capacities to digest alternative sources of carbohydrates and are at lower risk of NEC than preterm pigs delivered at 92% of term and fed formulas with MDX. An important consideration is that preterm infants at <30 weeks have limited brush border membrane carbohydrase activities (maltodextrinase has not been measured) and the responses to formulas or fortifiers with different sources of carbohydrate have not been adequately studied.

Healthy preterm pigs have total intestinal maltodextrinase capacities capable of hydrolyzing approximately 0.9 g · kg⁻¹ · day⁻¹ (our unpublished data), or approximately 15% of the dietary load provided to the MTX pigs that had a 50% incidence of NEC. Decreasing the MDX fraction to 15% of the carbohydrate (0.9 g · kg⁻¹ · day⁻¹) reduces the incidence of NEC among preterm pigs, but still does not prevent NEC ⁽¹⁰⁾.

Lactose Does Not Cause Necrotizing Enterocolitis

The amount of LAC in the LAC formula is approximately 2-fold what is present in COL from sows and cows and at the volume fed resulted in a dietary load of 6 g · kg⁻¹ · day⁻¹. This exceeds the estimated capacities of preterm pigs harvested at the same gestational age to digest approximately 2 to 3 g of LACkg⁻¹ · day⁻¹⁽¹⁰⁾, which is matched with the 3 g · kg⁻¹ · day⁻¹ of LAC present in the bovine COL fed to the COL pigs ⁽²¹⁾. Yet, the discrepancy between lactase capacities and dietary load of LAC was not associated with NEC or any other obvious health challenge. The limited lactase activities of preterm infants, incomplete digestion of LAC, higher breath hydrogen, and runny stools led to the concept of LAC malabsorption ⁽²²⁾. Even term infants have incomplete LAC digestion ⁽²³⁾ and there is no evidence to suggest that LAC or LAC malabsorption is a risk factor for NEC or other gastrointestinal pathologies. Moreover, treatment of formula with lactase to compensate for the LAC malabsorption does not improve growth of preterm infants or provide any significant benefit ⁽²⁴⁾.

How Might Maltodextrin Increase Necrotizing Enterocolitis Risk?

Even though the microbiomes of preterm pigs and preterm infants delivered by caesarian section are likely to differ, there is no evidence linking NEC with 1 or more specific bacteria ⁽²⁵⁾. Instead, the restriction of NEC to pigs fed the formula with MDX suggests altered bacterial metabolism in response to the 2 different carbohydrate sources is a more likely contributor to NEC risk. Bacterial fermentation produces gases and metabolites, including short-chain fatty acids, with the products varying among sources of carbohydrate ⁽²⁶⁾. Although breath hydrogen is lower when preterm infants are fed formula with partial replacement of LAC with MDX ⁽²³⁾ the health benefit is uncertain. In fact, the higher breath hydrogen of preterm infants fed breast milk is not associated with increased risk of NEC, is comparable to when formula with only LAC is fed, and has fostered speculation that undigested LAC is a prebiotic that accelerates the establishment of a commensal microbiome ⁽²⁷⁾. However, metabolites resulting from excessive LAC fermentation in conjunction with the presence of certain bacteria can be detrimental, as exemplified by studies with gnotobiotic quail ⁽²⁸⁾. There is a need to better understand how the preterm gut responds to changes in the absolute and relative concentrations of short-chain fatty acids and other bacterial metabolites resulting from fermentation of different undigested carbohydrates and other ingredients.

Growth Responses

We used bovine COL as the positive control because of availability, proven protection against NEC ^(10,11), the potential to promote growth ⁽²⁹⁾, and the possible use for human preterm infants (https://www.clinicaltrials.gov/ct2/show/NCT02054091). Based on the reported composition of bovine COL ⁽²¹⁾ the COL pigs were provided about 156 kcal · kg⁻¹ · day⁻¹, with (in g · kg⁻¹ · day⁻¹) protein = 17.9, fat = 8.0, and lactose = 3. These values are slightly higher than the 15.9, 5.1, and 2.8 g · kg⁻¹ · day⁻¹ of protein, fat, and lactose if the pigs had been fed sow COL collected during the first 12 hours after farrowing ⁽³⁰⁾. Although preterm pigs fed bovine COL are protected against NEC (present study; 11), the lack of growth during the 6 to 7 days of EN was unexpected because it contrasts with the growth of preterm pigs when the volume of COL is gradually increased ⁽³¹⁾, but does coincide with the abundance of indigestible protein ⁽³²⁾ and the lower fat and carbohydrate intake compared with the pigs fed the 2 formulas.

Neither of the 2 formulas resulted in growth that was comparable to the 5% to 6% gain per day of pig fetuses during a comparable period of gestation ⁽³³⁾ or to the rapid growth of suckling pigs. This is attributed to the increased energy expenditure associated with extrauterine life and to the volume of formula fed being approximately 50% of what term pigs consume ⁽³⁴⁾. However, when we attempted to feed preterm pigs with 250 mL · kg⁻¹ · day⁻¹ of the LAC-based formula, this caused >70% morbidity and mortality (unpublished data). The slightly lower growth for the MDX compared with the LAC pigs corresponds with the presence of undetected subclinical NEC.

Intestinal Responses to Different Enteral Nutrition Support

Preterm birth compromises the rapid intestinal growth typical of late gestation ⁽³⁵⁾, particularly if PN is required ⁽¹¹⁾, and the immature gastrointestinal tract has limited abilities to process dietary inputs. The increased size and protein synthesis of the small intestine in response to COL (present study ^(36,37)) along with the resistance to NEC are of obvious benefit for preterm infants.

Although actual dietary loads of glucose did not differ markedly among the 3 treatments, the fetal and preterm intestine has minimal abilities to adaptively modulate nutrient uptake to match changes in diet composition ⁽¹¹⁾. Moreover, the similar uptake capacities hence total intestinal densities of glucose transporters for pigs in the 3 treatments suggest COL stimulates intestinal growth, but without increasing the expression of glucose transporters, explaining the lower rates of glucose uptake per milligram tissue for COL relative to LAC and healthy MDX pigs. Furthermore, the glucose uptake capacities for LAC and MDX pigs exceeded the 6 g of aldohexose that would have been available if the LAC and MDX fed during a 24-hour period was completely hydrolyzed (glucose only for MDX and a combination of glucose and galactose for LAC), and for the aldohexose available from the LAC in cow COL.

Weights of Other Organs

Because brain growth is conserved at the expense of other tissues ⁽³⁸⁾, absolute brain weight is consistent across a wide range of body weights, decreasing only with very low birth weights. Therefore, the smaller brains of the COL pigs are significant and surprising because this suggests chronic feeding of exclusively COL may compromise brain growth. The smaller spleens of the COL pigs, even when normalized to body weight, suggests COL reduces infection and immune responses, although this was not directly assessed in the pigs. The similar absolute weights for the hearts and lungs among pigs in the 3 treatments indicate growth of these tissues was conserved and as a consequence, the hearts and lungs of the smaller COL pigs were proportionally larger (g/kg body weight) compared with the LAC and MDX pigs. The kidneys, liver, and pancreas grew in proportion with body weight gains.

Serum Chemistries

The relationship between BUN and protein intake during infancy ⁽³⁹⁾ is based on the extent of amino acid oxidation and is an indicator of both protein intake and quality. The higher BUN values for the COL pigs indicate a higher proportion of the protein was oxidized rather than accreted as tissue protein, which is consistent with the lack of growth. In contrast, the lower BUN values for the LAC and MDX pigs suggest a larger fraction of the protein was used for growth rather than oxidized and are comparable to values for term pigs fed larger volumes of low protein formulas with 15 and 30 g protein/L ⁽⁴⁰⁾ and for preterm infants fed a low protein formula that still promotes growth ⁽⁴¹⁾. Although body composition was not determined, the lower BUN values for pigs fed the formulas suggest a greater proportion of dietary protein was partitioned into lean body mass rather than the amino acids being deaminated and converted to stored fat.

The serum chemistries revealed other features pertaining to the chronic feeding of COL. The higher sodium and chloride of the COL pigs (21% and 26% higher compared with LAC and MDX pigs) correspond with the higher sodium and chloride concentrations in bovine COL compared with mature milk (35% and 31%, respectively ⁽⁴²⁾). Although electrolyte composition of the formula was not determined, the discrepancies between COL and mature milk may be an adaptive mechanism to assist in the transition of fetuses to the extrauterine environment. The higher serum lipids of the COL pigs exceeded the difference between the fat content of COL and the formulas, suggesting COL enhances hepatic lipogenesis and lipoprotein secretion. The lower serum lipids and higher aspartate aminotransferase activities measured in the MDX pigs with NEC are consistent with liver dysfunction.

Perspectives

The differential NEC incidence between pigs fed formulas with either LAC or different amount of MDX (present study, 10) provides an opportunity to explore the etiology, pathogenesis, and pathophysiology of NEC, which have remained uncertain, despite decades of research ⁽⁴³⁾. Furthermore, a formula that does not induce NEC will facilitate the evaluation of ingredients without NEC as a confounding factor. The present findings demonstrate the preterm pig model provides opportunities to evaluate the efficacy and safety of other carbohydrates and novel ingredients for use in preterm formula. This includes defining the proportion and duration of EN that can be represented by COL to obtain benefits without compromising growth and outcomes.