Nutrition and Metabolism Research Oral Paper Session Abstracts

Abstract

Sunday, March 23, 2025

SU30 Parenteral Nutrition Therapy

SU31 Enteral Nutrition Therapy

SU32 Malnutrition and Nutrition Assessment

SU33 Critical Care and Critical Health Issues

SU34 GI, Obesity, Metabolic, and Other Nutrition Related Concepts

SU35 Pediatric, Neonatal, Pregnancy, and Lactation

Parenteral Nutrition Therapy

Abstract of Distinction

Shaurya Mehta, BS¹; Chandrashekhara Manithody, PhD¹; Arun Verma, MD¹; Christine Denton¹; Kento Kurashima, MD, PhD¹; Jordyn Wray¹; Ashlesha Bagwe, MD¹; Sree Kolli¹; Marzena Swiderska-Syn¹; Miguel Guzman, MD¹; Sherri Besmer, MD¹; Sonali Jain, MD¹; Matthew Mchale, MD¹; John Long, DVM¹; Chelsea Hutchinson, MD¹; Aaron Ericsson, DVM, PhD²; Ajay Jain, MD, DNB, MHA¹

¹Saint Louis University, St. Louis, MO; ²University of Missouri, Columbia, MO

Financial Support: None Reported.

Background: Total parenteral nutrition (TPN) provides lifesaving nutritional support intravenously, however, is associated with significant side effects. Given gut microbial alterations noted with TPN, we hypothesized transferring intestinal microbiota from healthy controls to those on TPN would restore gut-systemic signaling and mitigate injury.

Methods: Using our novel ambulatory model (US Provisional Patent: US 63/136,165), 31 piglets were randomly allocated to enteral nutrition (EN), TPN only, TPN + antibiotics (TPN-A) or TPN + post pyloric intestinal microbiota transplant (TPN-IMT) for 14 days. Gut, liver, and serum samples were assessed though histology, biochemistry, and qPCR. Stool samples underwent 16s rRNA sequencing. PERMANOVA, Jaccard and Bray-Curtis metrics were performed.

Results: Significant bilirubin elevation in TPN and TPN-A vs EN (p < 0.0001) was prevented with IMT. Serum cytokine profiles revealed significantly higher IFN-G, TNF-alpha, IL-beta, IL-8, in TPN (p = 0.009/0.001/0.043/0.011), with preservation upon IMT. Significant gut-atrophy by villous/crypt ratio in TPN (p < 0.0001) and TPN-A (p = 0.0001) vs EN was prevented by IMT (p = 0.426 vs EN). Microbiota profiles using Principal Coordinate Analysis (PCA) demonstrated significant overlap between IMT and EN, with the largest separation in TPN-A followed by TPN, driven primarily by firmicutes and fusobacteria. TPN altered gut barrier (Claudin-3 and Occludin) was preserved upon IMT. Gene expression showed upregulation of CYP7A1 and BSEP in TPN and TPN-A, with downregulation of FGFR4, EGF, FXR and TGR5 vs EN and prevention with IMT. In a subgroup analysis on TPN and EN, regional gut integrity differences were analyzed through the varying presence of E-cadherin and Occludin in the segments of proximal gut, distal gut, and the colon. E-Cadherin and Occludin levels were significantly decreased in the TPN only group as compared to the EN group.

Conclusion: This work provides novel evidence of gut atrophy, liver injury, gut barrier dysfunction and microbial dysbiosis prevention with post pyloric IMT, challenging current paradigms into TPN injury mechanisms and further underscores importance of gut microbes as prime targets for therapeutics and drug discovery.

Dejan Micic, MD¹; Ena Muhic, MD²; Samuel Kocoshis, MD³; Loris Pironi, MD⁴; Simon Lal, MD, PhD, FRCP⁵; Farooq Rahman, MD⁶; Alan Buchman, MD, MSPH⁷; Jacqueline Zummo, PhD, MBA, MPH⁸; Khushboo Belani, MPH⁸; Eppie Brown, RN, MA⁸; McKenna Metcalf, BA⁸; Andrea DiFiglia, MS⁸; Palle Jeppesen, MD, PhD⁹; Jenny Han, MS¹⁰

¹University of Chicago, Chicago, IL; ²Rigshospitalet, Copenhagen, Sjelland; ³Cincinnati Children's Hospital Medical Center, Cincinnati, OH; ⁴University of Bologna, Italy, Bologna, Emilia-Romagna; ⁵Salford Royal NHS Foundation Trust, Salford, England; ⁶University College London, London, England; ⁷University of Illinois at Chicago, Glencoe, IL; ⁸Protara Therapeutics, New York, NY; ⁹Department of Intestinal Failure and Liver Diseases, Rigshospitalet, Copenhagen, Hovedstaden, Denmark; ¹⁰Pharmapace, a subsidiary of Wuxi AppTec, San Diego, CA

Financial Support: Protara Therapeutics.

Background: Choline is a quaternary amine that is an essential dietary nutrient in humans. It is essential for patients with intestinal failure (IF) who are dependent on Parenteral Support (PS), given that deficiency can lead to hepatic injury, neuropsychological impairment, muscle damage, and thrombotic abnormalities. Currently, there are no approved intravenous choline products for PS patients globally.

Methods: THRIVE-1 was a prospective, multi-center, cross-sectional, observational study to assess the prevalence of choline deficiency and liver injury in adolescents ( ≥ 12 years of age) and adult patients ( ≥ 18 years of age) with IF who are dependent on PS; (defined as at least 4 days/week on PS for at 10 to 24 weeks [capped at 25%] or 24 weeks or longer). Data collection occurred during a single clinic visit.

Results: Of 78 enrolled patients, 55% were male, 92% White, and 96% Not Hispanic or Latino. Mean age was 52 years (SD: 16.6), and BMI was 23.0 kg/m² (SD: 3.8; Table 1). Patients had received PS for a mean duration of ~9 years and a mean PS frequency of 6.6 days/week (SD: 0.9). Most patients received mixed (40%, 31/78) or plant-based lipids (49%, 38/78; Table 1). Patients had at least one of the following underlying conditions based on ESPEN Pathophysiological IF Classification: 59% (46/78) Short Bowel Syndrome, 46% (36/78) Mucosal Diseases, 33% (26/78) Chronic Intestinal Dysmotility Disorders, 8% (6/78) Mechanical Obstruction, and 6% (5/78) Intestinal Fistulae. Choline deficiency, defined as a plasma free choline concentration <9.5 nmol/mL, was present in 78% (61/78) of patients, with a mean plasma free choline concentration of 7.5 nmol/mL (SD: 3.9) ranging from 2.6 to 27.1 nmol/mL (Table 2). The prevalence of choline deficiency was also evaluated by age group, PS duration, lipid type and underlying condition (Table 2). Among the choline-deficient participants, 63% (38/60) had liver injury (defined as any elevated liver tests [>1.5*ULN; ALP, AST, ALT, GGT, Direct Bilirubin, Total Bilirubin] or steatosis [MRI-PDFF ≥ 8%]; Table 2).

Conclusion: The high prevalence of choline deficiency among patients with IF who are dependent on PS (overall, and across age groups, and regardless of PS duration, lipid type or underlying condition) emphasizes the need for choline supplementation and underscores the need for IV choline availability for this patient population to address an unmet need. Significant heterogeneity of liver injury was observed and warrants further investigation. Choline Chloride for injection (IV Choline Chloride), a phospholipid substrate replacement therapy, is being developed as a source of choline for long-term PS-dependent patients.

Table 1. Overview of Demographics, Baseline Characteristics, and Overview of Parenteral Support History.

Table 2. Overview of Choline Deficiency and Liver Injury.

Note: Percentages are based on the number of patients in the Enrolled Set with observed data.

Note: Choline deficiency is defined as <9.5nmol/ml. Various studies in PS-dependent patients report choline deficiency as baseline concentrations of plasma free choline ranging from approximately 5.2 ± 2.1 nmol/mL to 7.15 ± 2.5 nmol/mL (Buchman et al., 1993; Buchman et al., 1994; Buchman et al., 2001a; Compher et al., 2002).

Note: IF Classification is based on ESPEN Pathophysiological IF Classification; Patients may fall into more than one category

Note: Liver injury was defined as any elevated liver tests (1.5xULN; ALP, AST, ALT, GGT, Direct Bilirubin, Total Bilirubin) or Steatosis (MRI-PDFF ≥8%).

Ismail Pinar, MD¹; Thor Nielsen, MSc²; Lise Soldbro, MD²; Mark Berner-Hansen, MD²; Palle Jeppesen, MD, PhD¹

¹Department of Intestinal Failure and Liver Diseases, Rigshospitalet, Copenhagen, Hovedstaden, Denmark; ²Zealand Pharma A/S, Copenhagen, Denmark, Soeborg, Hovedstaden

Financial Support: Zealand Pharma A/S.

Background: Glucagon-like peptide-2 (GLP-2) is a neuroendocrine intestinal hormone, which facilitates intestinal adaptation and absorptive capacity. Glepaglutide is a long-acting GLP-2 analog in late-stage development for treatment of short bowel syndrome (SBS) patients. We aimed to assess the efficacy of glepaglutide on intestinal absorption, parenteral support use and safety in SBS patients.

Methods: EASE (Efficacy and Safety Evaluation) SBS-4 is a single-center, open-label, one-arm phase 3b trial. Patients received once-weekly 10 mg glepaglutide subcutaneous dose via a ready-to-use autoinjector. Adjustments to PS volume were guided by a pre-specified algorithm. For the first 24 weeks, patients adhered to a fixed drinking menu. Thereafter, beverage consumption was unrestricted intending to simulate habitual conditions. 48-hour metabolic balance studies were performed at baseline and week 24. Percentage changes (mean ± SD, p value) from baseline to weeks 24 and 52 in PS volume, contents and nutrient absorption (wet weight, electrolytes, energy by bomb calorimetry and macronutrients (mean, p value)) were evaluated by paired t-tests.

Results: Ten adult SBS patients were included: mean age 55 years, 5 females, 8 with intestinal failure and 8 without colon-in-continuity. Mean increase from baseline in wet weight absorption was 56% (398 ± 582g/day, p = 0.06). When adjusting for baseline PS volume, in a post-hoc analysis, the results were statistically significant (p = 0.04). Mean increase in energy absorption was 23% (1038 ± 1182 kJ/day, p = 0.02). Post-hoc analysis showed a statistically significant increase in carbohydrate absorption by 40 ± 48 g/day (p = 0.03), while absorption of proteins, lipids, sodium, potassium, calcium numerically increased. Baseline mean PS volume was 2.55 ± 1.72 L/day and decreased by -25% (-0.76 L/day, p = 0.03) at week 24 and by -30% (-0.80 L/day, p = 0.01) at week 52. PS mean energy content decreased by -37% at week 24 (-859 kJ/day, p = 0.004) and remained stable at week 52. PS carbohydrate content was reduced by -42% (-702 kJ/day, p = 0.008) and remained low at week 52. PS protein content decreased by -36% (-151 kJ/day, p = 0.02) at week 24, while lipid content decreased by -26% (-81 kJ/day, p = 0.32) at week 52. Body weight remained unchanged at weeks 24 and 52. Treatment was assessed to be safe and well-tolerated.

Conclusion: Once weekly glepaglutide treatment resulted in clinically relevant increases in intestinal absorption of wet weight, energy, electrolytes, and macronutrients at 24 weeks. Furthermore, sustained reductions in PS volume and nutrient needs over 52 weeks were achieved while maintaining body weight and hydration. Treatment was assessed to be safe and well-tolerated. The results support glepaglutide as a potentially new long-acting GLP-2 analog for the management of SBS patients.

Nasiha Rahim, MS, DO¹; Caitlin Jordan, MS, RDN, LDN, CNSC¹; Macy Mears, MS, RDN, LDN, CNSC¹; Mary Graham, RD, CSP, LDN, CNSC¹; Aubrey Sanford, MS, RD, LDN¹; Timothy Sentongo, MD¹

¹University of Chicago, Chicago, IL

Financial Support: None Reported.

Background: Newborn infants born with very low birth weight (VLBW) <1500 g and prescribed parenteral nutrition (PN) are at risk for developing hypophosphatemia from inadequate intake or refeeding syndrome [1]. However, the dosing protocol for calcium (Ca) and phosphorous (PO4) during TPN in newborn infants is primarily directed toward optimizing bone mineralization, which is an important long-term goal. Our TPN team observed that during the first 7-days of life, there is frequent forced deviation from the A.S.P.E.N 2014 clinical guidelines that recommends a Ca:PO4 ratio of 1.7:1 (mg: mg) or 1.3:1 (mmol: mmol) in short-term PN in neonates [2]. Therefore, the purpose of this quality improvement (QI) project was to 1) determine the frequency of deviating from in-ratio dosing of Ca:P because of hypophosphatemia during the first seven days of PN in VLBW; 2) Identify the clinical factors associated with increased risk for developing hypophosphatemia during the first seven days of life. The long-term objective was to pre-emptively dose Ca:PO4 in infants at high risk for hypophosphatemia.

Methods: This was an institution-approved Quality Improvement (QI) project to identify the risk factors for hypophosphatemia during PN nutrition support of newborn infants. As per institutional protocol, every newborn infant had serum electrolytes, Ca and PO4, measured at baseline, and had regular monitoring as the PN components were adjusted to meet nutritional goals, including in-ratio dosing of Ca:P. The reference range for normal serum PO4 was defined as 5.0 to 7.3 mg/dL. The outcome was the frequency of deviating from in-ratio dosing of Ca:PO4 because of hypo- or hyperphosphatemia. Data collection for the QI lasted six months (7/1/2023 to 12/31/2023). Eligibility included birthweight ≤1500 g, central venous access availability, and PN initiation at birth (day of life ‘0’). Data was collected from the initiation of PN at birth to five days. The information included gestation age, birth weight, sex, serum Ca, PO4, and daily PN composition. The outcome was the frequency of deviating from guideline-based in-ratio dosing of Ca:P of 1.7:1 (mg: mg) or 1.3:1 (mmol: mmol) because of hypo- or hyperphosphatemia.

Results: Sixty-four infants (58% males) were enrolled. The birth weight (median/IQR) was 0.987 (0.718,1.25) kg, and their median gestation age was 27.5 (25, 30) weeks. The prevalence of hypophosphatemia was 66.7%, 63.3%, 76.4%, 76.6%, 73.9%, and 52.8%, and days ‘0’, 1, 2, 3, 4, and 5 of life, respectively. However, only 16 infants had birth weight percentiles <10%, corresponding to small for gestation age, and thus at increased risk of developing refeeding hypophosphatemia, a condition that occurs when malnourished individuals, such as premature infants, are fed and their metabolic rate increases, leading to a rapid drop in serum phosphate levels. The frequency of deviating from in-ratio dosing of Ca:PO4 because of serum hypophosphatemia increased from 7.8% on day ‘0’ to 29.7%, 58.1%, 73.8%, 79.3% and 83.3% on days 0, 1, 2, 3, 4 and 5 of life respectively, as seen in Figure 1.

Conclusion: During the first week of PN therapy in infants with VLBW, in-ratio dosing of Ca:P was associated with a high prevalence of serum hypophosphatemia regardless of birth weight percentiles. These findings highlight the urgent need for a more nuanced approach to dosing Ca:PO4 in VLBW infants. Applying the current ASPEN guidelines for in-ratio dosing of Ca:PO4 appears to be suboptimal for maintaining normal serum phosphate status in VLBW infants during the first week of life. Therefore, during the first week of PN therapy, neonatal practitioners should be prepared to revert from in-ratio-based Ca:PO4 to serum-level-directed dosing of Ca and PO4 in VLBW infants.

Depiction of calcium and phosphorus use in parenteral nutrition added in ratio, higher than ratio, lower than ratio, and no additions on each day of life.

Figure 1. Parenteral Nutrition Addition of Calcium:Phosphorus Ratio (%).

Best of ASPEN-Parenteral Nutrition Therapy

Abstract of Distinction

Thanaphong Phongpreecha¹; Marc Ghanem¹; Jonathan Reiss²; Tomiko Oskotsky³; Samson Mataraso¹; Taryn Ng²; Boris Oskotsky³; Jacquelyn Roger³; Jean Costello³; Steven Levitte⁴; Brice Gaudillière¹; Martin Angst¹; Thomas Montine¹; John Kerner¹; Roberta Keller³; Karl Sylvester¹; Janene Fuerch¹; Valerie Chock¹; Shabnam Gaskari²; David Stevenson¹; Marina Sirota³; Lawrence Prince¹; Nima Aghaeepour, PhD¹

¹Stanford University, Stanford, CA; ²Lucile Packard Children's Hospital, Stanford, CA; ³University of California, San Francisco, CA; ⁴Stanford Physician, Stanford, CA

Financial Support: We would like to acknowledge contributions from other authors at Stanford University that could not be included due to authorship limit. This work was supported by the NIH grant R35GM138353, NCATS UL1TR001872, NICHD R42HD115517, the Burroughs Wellcome Fund (1019816), the March of Dimes, the Alfred E. Mann Foundation, the Stanford Maternal and Child Health Research Institute through Stanford's SPARK Translational Research Program, Stanford High Impact Technology (HIT) Fund, and Stanford Biodesign.

Background: Prematurity is a leading cause of infant mortality, with survivors often facing long-term complications. Total Parenteral Nutrition (TPN) is essential for supporting neonates, especially preterm infants, who cannot tolerate enteral feeds. However, TPN formulation is highly complex, prone to errors, and varies significantly across institutions, leading to inconsistent and sometimes suboptimal care.

Methods: In response to these challenges, we developed TPN2.0, an AI-driven model that optimizes and standardizes TPN prescriptions using a data-driven approach based on existing electronic health records (EHRs; Figure 1). TPN2.0 was trained on a decade of TPN prescriptions from 2011 to 2022, covering 79,790 orders from 5,913 patients at Lucile Packard Children's Hospital at Stanford. The model was then validated on an independent dataset from UCSF Benioff Children's Hospital, consisting of 63,273 TPN orders from 3,417 patients from 2012 to 2024. By combining advanced machine learning techniques, TPN2.0 developed 15 standardized TPN formulations. These formulas can then be personalized for each patient based on clinical characteristics such as lab values and demographic information, providing the benefits of standardization while maintaining precision for individual needs.

Results: In comparative analyses, TPN2.0 demonstrated high accuracy, with a Pearson correlation of R = 0.94 when compared with expert-designed TPN orders (Figure 2). In the external validation dataset at UCSF, the model achieved Pearson's R = 0.91, underscoring its generalizability across institutions. In a blinded study with healthcare providers on 192 comparisons, TPN2.0 recommendations received 53% higher rating scores than traditional TPN orders (Figure 3). This preference demonstrates the model's ability to generate clinically relevant, reliable TPN formulations. TPN2.0 was also associated with improved clinical outcomes. For example, in a retrospective analysis, patients whose prescribed TPN formula aligned with TPN2.0 saw a 5-fold reduction in the odds ratios (OR) of mortality compared to those whose prescriptions deviated from TPN2.0 recommendations. Other OR reduction includes 3 folds in necrotizing enterocolitis (p = 0.0007) and almost 5 folds in cholestasis (p = 4.29 × 10⁻³²; Figure 4). TPN2.0 also feature a physician-in-the-loop system, allowing clinicians to modify recommendations based on patient-specific conditions. In fact, our retrospective analysis showed that if physicians were to override TPN2.0 recommendations 20% of the times, the model correlation performance for next day recommendations increased by an average of 18%. This demonstrates adaptability to real-world clinical workflows while improving efficiency and reducing variability in TPN formulation.

Conclusion: TPN2.0 bridges the gap between individualized, error-prone TPN and overly rigid standardized TPN by offering personalized, scalable nutrition solutions. With its potential for centralized production and distribution, TPN2.0 is especially impactful for low- and middle-income countries (LMICs), where access to customized TPN is limited, offering a pathway to improved neonatal care in resource-constrained environments.

Figure 1. Process Comparison Between Current Practice and With TPN2.0.

Figure 2. TPN2.0's Correlation With Expert Practices.

(a) Multidisciplinary healthcare team members reviewed historical data and rated three TPN solutions from 0 to 100. A score of 0 indicated complete disagreement, while 100 meant the solution matched exactly what they would have prescribed. (b) From a total of 192 comparisons, TPN2.0 received higher rating scores than the TPN actually prescribed (M.W.U. P < 0.0001).

Figure 3. Blinded Validation Study Design and Results.

The odds ratios of various outcomes between TPN prescriptions that deviated from TPN2.0 recommendations (case) vs. TPN prescriptions that are similar to TPN2.0 (control).

Figure 4. Deviation From TPN2.0 Increases the Risk of Adverse Outcomes.

Trainee Award

Gulisudumu Maitiabula, PhD¹; Xinying Wang²

¹Jinling Hospital, Nanjing, Jiangsu; ²Wang, Department of General Surgery, Jinling Hospital, Medical School of Nanjing University, Nanjing, Jiangsu

Financial Support: None Reported.

Background: Intestinal failure (IF) is a clinically common and potentially fatal organ injury, and parenteral nutrition (PN)-associated liver disease (PNALD), a serious complication of long-term PN in IF patients, has been identified as a negative predictor of survival in IF patients. Previous studies have reported that the occurrence of PNALD is closely related to the changes of intestinal microbiota, but the changes of intestinal fungi and their effects on PNALD are still unclear. The aim of this study was to identify the intestinal fungal changes in PNALD patients and the specific mechanism of fungus-related PAMPs in the development of PNALD.

Methods: The fecal smear results of patients with IF were retrospectively analyzed. Fecal samples were prospectively collected from patients with IF for ITS sequencing and untargeted metabolomics analysis. A mouse model of PNALD was induced by total parenteral nutrition (TPN). Liver single-cell sequencing and proteomics were used to determine the reprogramming of liver macrophages during PNALD. Fecal metagenomics and single fungal intervention were used to further verify the role of gut fungi in PNALD.

Results: Fungal spores were found in 35.74% of IF patients, and the incidence of fungal spores in PNALD patients was significantly higher than that in non-PNALD patients. There was a significant difference in the β-diversity of fungi between the PN mice and the PN mice, and the g_Candida was significantly increased in the PN mice, indicating that PN caused changes in the abundance of intestinal fungi. ITS sequencing, liver single cell sequencing and single fungal colonization revealed that the intestinal fungi of PNALD patients were changed, and Candida tropicalis increased and mannan released was increased. The expression of C-type lectin receptors (CLRs) and related proteins in the liver of PNALD mice was significantly increased. These proteins include Mincle (Clec4e), Dectin-3 (Clec4d), Dectin-2 (Clec4n) and its downstream proteins Fcrγ (Fcer1g), Syk (Syk) and caspase recruitment domain-containing protein 9 (Card9). Lipid-associated macrophages (LAMs) related genes and specific markers of non-metastatic melanoma protein B (GPNMB) are significantly increased in the liver of PNALD mice and PNALD patients. The single fungi-Candida tropicalis intervention significantly aggravated the occurrence of PNALD.

Conclusion: Our results demonstrated that intestinal fungal changes and liver macrophage reprogramming were the key factors in the development of PNALD. Mannan derived from Candida tropicalis combined with CLRs activated LAMs to secrete GPNMB, thereby contributing to the development of PNALD. This study will provide new therapeutic targets and theoretical basis for the prevention and treatment of PNALD, which has good translational medicine value.

Table 1. Characteristics of ITS Participants.

Table 2. Characteristics of Metabonomic Participants.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Enteral Nutrition Therapy

Jenny Lee, MS, RD, LD, CNSC¹; Jenna Holder, MS, RD, LD, CNSC¹; Robyn Brown, MBA, RD, LD, CSSBB¹; Amanda L'italien, MS, RD, LD, CNSC¹; Benjamin Brofman, RN, BSN¹; Corrin Doming, RN, BSN¹; Georgia Fabbrini, MS, CCC-SLP¹; Germaine Ngog, RN, BSN, SCRN, PCCN, SCRN, ASLS¹; Karla Medina, RD, LD, CNSC¹; Jessica Carodine, MA, CCC-SLP¹; Judy Harrelson²; Tondi Martin, RN, BSN¹; Huimahn Choi, MD, MS³

¹Memorial Hermann Hospital - Texas Medical Center, Houston, TX; ²Memorial Hermann Hospital - TMC, Houston, TX; ³UT Health Houston McGovern Medical School, Houston, TX

Financial Support: None Reported.

Background: Enteral nutrition is the preferred method of feeding when patients are unable to consume adequate nutrition on their own and when they have functional and accessible guts. Nurses place most feeding tubes blindly at the bedside. However, a blind-placement method comes with four risks. First, when a post-pyloric feeding is required, the success rate of placement is very low, especially in patients with decreased gastrointestinal motility and abnormal gastrointestinal anatomy. Second, these patients often quickly accumulate a deficit of >10,000 calories while waiting for post-pyloric feeding. Third, 1.6% of all blind tube placements are inserted into the lungs, and 0.5% cause a major complication, including pneumothorax and death (Taylor, 2022). Four, multiple abdominal x-rays are taken for placement confirmation. To tackle these deficiencies, a multidisciplinary team evaluated the safety, accuracy, and effectiveness of an electromagnetic-guided (EMG) technology to guide feeding tube placement at bedside.

Methods: In summer 2019, a multi-disciplinary committee determined that an EMG device might improve feeding tube placements at bedside. Thirty EMG tubes were placed during a 7-day pilot in winter 2020. Tube locations were 0% lung, 10% gastric, and 90% small bowel. In fall 2021, nine superusers consisting of dietitians, nurses, and speech pathologists used one machine to start placing tubes Monday-Friday, 8am-4pm, after extensive training and competency tests. To prioritize high-risk patients, tube placements were limited to patients who required post-pyloric feeding tubes or had difficult bedside placements. However, this practice created some concerns. First, one machine did not meet the demand from a 800-bed hospital, including many with dysphagia. Second, superusers were too busy with their regular assignments. Lastly, approximately 40% of requests were placed outside the operating hours. In the winter of 2022, at the request of physician partners, the hours were extended to 7 days a week and two full-time positions were added to cover demand. In 2023, to further mitigate safety concerns, the team implemented a two-step method to verify airway placement and confirm esophageal insertion before advancing the tube in high-risk cases.

Results: The use of an EMG device to place tubes improves nutrition delivery by reducing the time to place and replace a feeding tube. From the 3319 EMG tubes placed between 2021 and 2023, the mean tube insertion time was 18 ± 22 minutes. Traditional blind placement takes much longer due to uncertainty in tube location. The accuracy of guided post-pyloric tube placement was high at 90%, which is similar to McCutcheon's 2018 study with 6290 placements. The small-bore feed tube (SBFT) location in that study was 87% post-pyloric, and the mean insertion time was 15 minutes. This technology improves safety compared with traditional method. X-ray confirmation of tube placement has been reduced by a striking 75%. The number of pneumothorax caused by tube placements has also significantly decreased between 2019 and 2023. The observed rate per 1000 case for iatrogenic pneumothorax rate was decreased from 0.33 (O/E ratio 1.2) in 2019 to 0.22 (O/E ratio 0.77) in 2022 and zero in 2023.

Conclusion: This study demonstrates that the EMG method and the small multi-disciplinary team approach are safe and reliable for placing SBFT at bedside. The strategy also enhances nutrition delivery by shortening the time to place and replace a feeding tube and improves patient safety by reducing the amount of x-ray exposure and incidents of lung placement.

Kami Benoit, DCN, RDN, LD, CNSC¹; Swarna Mandali, PhD, RD²; John Dumot, DO, FASGE³

¹Cleveland Clinic, Hudson, OH; ²Kansas University Medical Center, Kansas City, KS; ³University Hospitals, Cleveland, OH

Financial Support: None Reported.

Background: Provision of enteral nutrition in hospitalized patients can be delayed for multiple reasons, including lack of adequate enteral access. While gastric feeding may be relatively easy to initiate, it is not always appropriate and post-pyloric feeding may be indicated. Bedside feeding tube systems can safely and effectively assist clinicians with placement of temporary feeding tubes into the small bowel. Electromagnetic assisted device (EMAD) systems and the image assisted device (IAD) system are designed to help achieve enteric access while reducing risk of adverse events, time to placement, time to initiation of feed, and cost per placement. To our knowledge, no studies compare the use of real-time IAD systems and EMAD for effectiveness of post-pyloric placement on first attempt. This study aimed to compare rates of post-pyloric tube placement on first attempt using the IAD and the EMAD systems in one academic hospital. Associations with post-pyloric placement rates and diagnosis category were also performed.

Methods: This retrospective, observational study was approved by the University Hospitals IRB. It included a review of small bore bedside feeding tube placements in patients who were admitted to one hospital. Confirmation of tube placement was made via direct examination of the electronic medical record. Data collected included the documented date, patient demographics, type of tube used for insertion, and radiologic reports of feeding tube tip position. A p-value of <0.05 was considered statistically significant.

Results: In all, data was collected on a total of 236 tube placements (Figure 1, Table 1). One hundred eight tubes were placed using the IAD, while 128 were placed via the EMAD. Of the 108 placements using the IAD, 13% (n = 14) were post-pyloric on first attempt, while 87% (n = 94) were gastric placements on first attempt. The percentage of post-pyloric placements of first attempt using the EMAD was 45.3% (n = 58) and 54.7% (n = 70) were gastric placements on first attempt. The data demonstrated a statistically significant higher post-pyloric placement rate on first attempt using the EMAD (p < 0.001) according to Pearson's Chi Square analysis (Figure 2). Additional analyses using a layered chi-square testing was conducted to demonstrate any significant relationship between tube type and tube placement when controlling for diagnoses (Table 2). In those admitted with a neurological diagnosis, significantly more (p < 0.001) post-pyloric tube placements were completed with the EMAD (47.5 %, n = 21) than the IAD (13.5%, n = 5).

Conclusion: A significantly higher rate of post-pyloric tube placements on first attempt using the EMAD bedside feeding tube system were seen compared to the IAD bedside feeding tube system. There were significantly more post-pyloric placements on first attempt in patients admitted with a neurologic diagnosis compared to other admission diagnoses. Data presented here create a baseline comparison of post-pyloric placement rates for two bedside feeding tube placement systems. Findings suggest implications for future comparison and consideration for institutions treating populations with increased need for temporary post-pyloric feeding tubes.

Table 1. Patient Demographics (n = 236).

IAD = image assisted device, EMAD = electromagnetic assisted device Other – included general medical diagnoses not otherwise specified such as overdose, intoxication, or sepsis.

Table 2. Post-Pyloric Placement Rates and Diagnosis Category.

Subscript a = Pearson Chi-square Subscript, b = Fisher's exact test; cells with expected count <5 * p < 0.05.

*Statistical significance p < 0.05.

Figure 1. Selection Flowchart.

Figure 2. Post-pyloric placement rates.

Paola Bregni, MS¹; Lingtak-Neander Chan, PharmD¹; Michelle Averill, PhD, RDN¹; Mari Mazon, MS, RDN, CD¹

¹University of Washington, Seattle, WA

Financial Support: None Reported.

Background: Drug-nutrient interactions (DNIs) can result in clinically relevant implications on therapeutic plans or health. Reported prevalence of DNIs ranges from 6-70%, implicating variation in the assessment and interpretation among clinicians. Contributing factors of the variance include limited nutritional expertise among clinicians, reliance on outdated and/or anecdotal evidence, and lack of consensus on objectively and accurately assessing DNIs. This oversight represents a critical gap in healthcare, as there is currently no standardized, evidence-based clinical tool to assess the probability of DNIs. We developed the Nutrient Drug Interaction Probability Scale (NDIPS), a novel tool to guide clinicians to systematically evaluate potential DNIs and determine their clinical relevance.

Methods: The NDIPS was modeled after the Drug Interaction Probability Scale and consists of ten questions addressing various aspects of the patient-specific factors related to the potential interaction. Each question is assigned a point value, and the sum of these points is tabulated to the final probability score, categorizing the DNI as highly probable, probable, possible, or doubtful. An internal validation was performed using 8 published DNI case reports after a comprehensive PubMed search. Each case report was evaluated using the NDIPS, with the expectation that the tool would generate a “highly probable” or “probable” NDIPS score. An external validation involving 6 practicing clinicians applied the NDIPS to a selected case report and provided feedback on the tool's consistency in question interpretation, practicality in the questionnaire's design, and areas for improvement.

Results: The NDIPS classified 37.5% of the case reports as "highly probable" and 25% as “probable”. This finding was unexpected considering that all reported the presence of a DNI. Further analysis revealed that certain questions had stronger predictive properties for the final probability score. Specifically, questions 5-10 (Table 1) have a more significant impact on the final probability score when compared to questions 1-4. Questions 5-10 delve into critical and patient-specific aspects such as the timing of the interaction, effects of dechallenge/rechallenge, and alternative causes. In contrast, questions 1-4 focus on previously published interaction reports, proposed mechanism of interaction, and subjective or objective evidence. While these factors are still important, their impact on the final probability score is less significant than the factors addressed in questions 5-10. Figure 1 displays cumulative trajectories for each DNI pair, while Figure 2 presents a boxplot of results stratified by both probability groups and question categories. Notably, questions 5-10 address aspects of DNIs that are frequently overlooked or omitted during standard clinical assessment procedures of DNIs in clinical practice and in several published case reports, enhancing the tool's potential to improve DNI assessment. The external validation process demonstrated consistency in clinician interpretations of the NDIPS. Overall, the NDIPS effectively categorized DNI cases, with higher scores correlating to more complete information. Incomplete data reduced predictability, emphasizing the importance of comprehensive reporting in assessing drug-nutrient interactions.

Conclusion: Our internal validation demonstrated effectiveness in distinguishing the probability of DNIs based on the evidence presented and clinical development. By providing a structured, comprehensive assessment process, the NDIPS supports personalized, evidence-based care for potential DNIs. This approach could enhance patient care by avoiding unwarranted dietary restrictions or medication changes founded only on theoretical DNI risks.

Table 1. Nutrient Drug Interaction Probability Scale (NDIPS).

Affirmative responses from each question will increase the total cumulative score of NDIPS.

Figure 1. Plot Line of Cumulative NDIPS Trajectory by Individual DNI Pairs.

Graphical representation of the variability in the results from the question group categories (questions 1-4 and questions 5-10) stratified by probability group. Box-and-whisker plots illustrate median values, 25th–75th percentiles (box) and 10th–90th percentiles (whiskers). Notably, questions 5-10 represent a higher impact on the final probability score when compared to questions 1-4.

Figure 2. Boxplot of NDIPS Results Stratified by Both Probability Groups and Question Category.

Abstract of Distinction

Osman Mohamed Elfadil, MBBS¹; Yash Patel, MBBS¹; Suhena Patel, MBBS¹; Danelle Johnson, MS, RDN¹; Ryan Hurt, MD, PhD¹; Manpreet Mundi, MD¹

¹Mayo Clinic, Rochester, MN

Financial Support: None Reported.

Background: Diagnosis of gastrointestinal dysmotility (GID) and its pragmatically defined subclasses, chronic intestinal pseudo-obstruction, and enteric dysmotility, can be challenging. Additionally, patients with GID often require specialized nutrition care involving short- or long-term home enteral nutrition (HEN) or parenteral nutrition (PN). Nutrition outcomes in this group of patients are understudied, with some data suggesting higher EN-related complications. This review aims to evaluate a single-center experience of HEN-dependent patients with GID from our prospectively maintained dataset.

Methods: A retrospective analysis of nutrition history and outcome of patients with GID who received HEN at our program with an initiation date between January 2018 and December 2023. HEN-related complications, including enteral feeding intolerance, tube-related, and metabolic complications, were compared to patients without GID who received HEN at our program for other indications.

Results: A total of 2855 patients are included in this analysis. Patients without GID (n = 2715; 95.1%) (mean age 61.9 ± 33.1 years; 39.5% female). Patients with GID (n = 140; 4.9%) (mean age 51 ± 21.2 years; 70.8% female; mean BMI at EN initiation 24.4 ± 6.6) (Table 1). Notably, more females were in the GID group. Additionally, patients with GID had a lower mean BMI at the initiation of EN and were relatively younger at the initiation of HEN (Table 1). In the analysis of complications, patients without GID had a higher prevalence of complications while on HEN (59.2% in the GID group vs. 43.6% in the non-GID group; p < 0.001). Moreover, they also had a higher prevalence of enteral feeding intolerance (EFI) and tube-related complications (p < 0.001) (Table 1). In further sub-analysis of the GID group, we also noted significant and clinically relevant differences between CIPO (n = 24; 17.1%) and ED (n = 116; 82.9%) subgroups. Patients with ED were predominantly females (68.1 in the ED group vs. 44.2% in the CIPO group; p < 0.001) and younger (mean age of 42.1 ± 14.9 years in the ED group vs. 68.7 ± 22 years in the CIPO group; p < 0.001). While not statistically significant, key differences in the incidence of complications are a higher incidence of EFI in the ED group and higher tube-related complications in the CIPO group (Table 2).

Conclusion: Our findings demonstrate that GID is clinically challenging in the HEN setting. Patients with gut motility disorders are less likely to tolerate EN and, therefore, fail to achieve nutrition goals and may require PN. Additionally, ED is the most prevalent type of GID that impacts younger and thinner females and is associated with a higher incidence of EFI. Patients with GID, therefore, may require closer follow-up, individualized HEN regimens with lower thresholds to interventions like transitioning to predigested enteral formulas.

Table 1. Clinical Characteristics and EN Tolerance.

Table 2. Nutrition Outcomes by Subclasses of GID (CIPO vs. ED).

Abstract of Distinction

Jan Powers, PhD, RN, CCNS, CCRN, NE-BC, FCCM, FAAN¹; Janette Richardson, MSN, RN, AGCNS-BC, CCRN²; Annette Bourgault, PhD, RN, CNL, FAAN³

¹Parkview Health, Westfield, IN; ²Parkview Health, Fort Wayne, IN; ³University of Central Florida, Orlando, FL

Financial Support: None Reported.

Background: Small bore feeding tubes (FTs) are frequently used to provide nutrition and medications to acutely ill patients. Distal FT location is important to minimize risk for aspiration and promote absorption of nutrients. Current guidelines call for routine FT placement verification, yet there are no valid methods available to perform this assessment at the bedside. FT migration occurs within the gastrointestinal (GI) tract due to normal GI motility, however little evidence exists on retrograde migration. Migration of the distal FT from the small bowel into the stomach or the esophagus may place patients at risk for microaspiration into the pulmonary system. It is unknown if retrograde FT migration may occur as a result of patient repositioning or routine procedures. The purpose of this study was to determine if small bore FTs migrate backwards (retrograde) in critically ill adults receiving usual care. The secondary aim was to identify factors that may affect FT migration. These findings may inform recommendations for clinical practice regarding the frequency for routine FT placement verification.

Methods: A longitudinal, repeated measures design was used with a convenience sample of 120 ICU patients. An electromagnetic placement device was used to verify daily FT position to assess for migration. Data were collected every 24 hours for up to 5 days while the FT was in place. Twenty variables included demographic data, clinical treatments, patient positioning, and procedures to assess for their effect on FT migration.

Results: Patient enrollment was completed in July 2024 (n = 120). Data analysis is currently in progress including bivariate descriptive characteristics of the study population and their association with FT migration status. Of the 120 patients, 61 had FTs for the full 5 days; 96 had FTs for 3 or more days. Differences in clinical and demographic characteristics between the groups will be assessed by chi-square statistics for categorical variables and two-sample t-tests (or Wilcoxon for nonparametric) for continuous variables. Preliminary data showed most FT migration occurred within the duodenum. We were unable to advance the stylet in 4 FTs that were clogged or kinked. Episodes of retrograde migration appeared to be mechanical related to patient agitation and/or pulling on the FT. Endotracheal tubes were removed in 32 patients; 94% had no FT migration as a result of the extubation. Further investigation is needed to determine the cause of 2 FTs that were removed during extubation; FT removal may have been intentional because the FTs were not replaced. One patient with a tracheostomy had no FT migration. An additional FT was intentionally removed during a procedure.

Conclusion: Preliminary data revealed interesting findings related to FT migration and activities such as extubation. Retrograde FT migration rarely occurred and was most likely caused by patient agitation and pulling on FTs vs. procedural or position changes. FTs typically remained in place during endotracheal tube extubation. Our results challenge the current recommendations to verify distal FT location every 4 hours. Final results and practice recommendations will be available for presentation at the conference.

Abstract of Distinction

Osman Mohamed Elfadil, MBBS¹; Christopher Staley, PhD²; Levi Teigen, PhD, RD³; Lisa Miller, RDN, LD, CNSC⁴; Lisa Epp, RDN, LD, CNSC, FASPEN⁴; Adele Pattinson, RDN¹; Danelle Johnson, MS, RDN¹; Yash Patel, MBBS¹; Ryan Hurt, MD, PhD¹; Manpreet Mundi, MD¹

¹Mayo Clinic, Rochester, MN; ²University of Minnesota, Minneapolis, MN; ³University of Minnesota, St. Paul, MN; ⁴Mayo Clinic Rochester, Rochester, MN

Financial Support: This research is supported by a Grant from Real Food Blends.

Background: Enteral nutrition (EN) is crucial for patients unable to maintain oral autonomy to prevent gut mucosal atrophy and maintain the gut barrier. Current guidelines recommend using a standard polymeric or high-protein standard formula in patients requiring EN. Dietary homogeneity, however, has been shown to negatively impact the health of the gut microbiome. Dietary diversity (i.e., a variety of foods across food groups per day) is associated with increased microbiome stability – a measure of microbiome health. Increasingly used, blenderized tube feeding (BTF) products with whole food ingredients may support increased dietary diversity and improved microbial health compared to standard formulas in patients requiring EN. This randomized prospective pilot study aims to examine the effects of EN on the gut microbiome using a standard versus commercial blenderized whole-food-based formula.

Methods: Consenting home EN dependent adults were recruited by invitation and randomized to switch to either a fiber-containing standard formula (SF) or a commercial whole-food-based BTF (BTF) for 4-6 weeks. We excluded those with diabetes mellitus, recent or current exposure to antibiotics, pre- and pro-biotics supplements, surgically altered gut anatomy, active malignancy, immunological conditions, or a history of organ transplantation. In addition to baseline demographic and clinical characteristics, we collected clinical nutrition variables. Stool samples were collected at baseline, a week (7-10 days) after switching to the study formula, and 4-6 weeks after switching to the study formula. Microbial communities were characterized by shotgun sequencing, and taxonomic and functional annotations were made using MetaPhlAn4.1 and HUMAnN3.9, respectively. Differences in abundances of genera between groups were evaluated by Kruskal-Wallis test and using MaAsLin3. Shannon and Bray-Curtis indices were used to estimate alpha and beta diversity, respectively.

Results: Nine patients with chronic dysphagia completed the study, including 5 in the BTF group (mean age 69.4 ± 8.6 years; 80% male) and 4 in the SF group (mean age 49.5 ± 24.8 years; 75% male). Shannon indices were similar between the study groups at all time points, and both groups had a comparable distribution of predominant species at baseline. Bacteroides uniformis and Faecalibacterium prausnitzii relative abundances increased across the 7-10 day and 4-6 week timepoints in the BTF group compared to the SF group (Figure 1). BTF supported B. uniformis more than SF at both the 7-10 day (4% in the BTF group vs. < 1% in the Standard group; p = 0.027) and 4-6 week timepoint (6% in the BTF group vs. 0% in the Standard group; p = 0.024) (Figure 2). On analysis of similarity, differences in community composition between both groups were observed (R = 0.39, P < 0.001). Over time, more similarity is observed in BTF compared to SF as microbiota directionally clustered towards beneficial bacteria such as B. uniformis (Figure 3).

Conclusion: These findings suggest a possible role for BTF in supporting the growth of beneficial gut microbiota species compared to SF in patients requiring EN. However, further work with a larger sample size is needed to validate these findings and further elucidate the effects of BTF on the gut microbiota.

Figure 1. Distribution of Predominant Species.

Figure 2. Distribution of Predominant Species.

Figure 3. Significant Difference in Community Composition Between Treatments.

Malnutrition and Nutrition Assessment

Kaylyn Koons, BS¹; Carley Rusch, PhD, RDN, LDN²; Anice Sabag-Daigle, PhD²; Wendy Dahl, PhD, RD¹

¹University of Florida, Gainesville, FL; ²Abbott Nutrition, Columbus, OH

Financial Support: This manuscript was not funded by any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. As employees of Abbott Nutrition, CR and ASD receive salaries for their professional responsibilities.

Background: Obesity is a highly prevalent chronic disease characterized by elevated fat accumulation and associated with increased risk of comorbidities such as diabetes and cardiovascular disease. Body mass index (BMI) is used as a screening tool to identify risk of obesity, and those with elevated BMIs and treatment will usually require behavior modification (diet/exercise), pharmacologic, and/or surgical interventions to induce an energy deficit for body weight reduction. The majority of adults do not meet the Adequate Intake for dietary fiber, and recent research has highlighted the role of dietary fiber supplementation as an adjunctive strategy for weight loss and body composition improvement. Supplementation of dietary fibers has been shown to influence satiety, glycemic control, and lipid metabolism in people with higher BMIs. These data suggest dietary fiber supplementation could be a valuable tool in the management of obesity. Therefore, this systematic review examines the association between dietary fiber supplementation and changes in body weight and body composition among adults with higher BMIs.

Methods: PRISMA guidelines for systematic reviews were followed, and two reviewers identified articles in MEDLINE/PubMed, EMBASE and the Cochrane Library databases published between January 1, 2009, and August 8, 2024. Randomized, controlled trials with 1) dietary fiber supplementation alone and in combination with other dietary interventions (e.g., probiotics) and/or added within foods for ≥ 4 weeks; 2) non-pregnant adults; 3) mean BMI ≥ 25; and 4) reported outcomes on body weight were included. Outcomes on waist circumference, body composition, and GI function were also reviewed when available.

Results: The search yielded 1406 unique records for screening, and 158 underwent full-text review. Sixty-one studies with a total of 3554 participants were included in the analysis. Ages ranged from 21-64 years, with mean baseline BMIs between 26-40. Among these 61 studies, fiber supplementation was characterized by using either prebiotic fibers (n = 33) or other dietary fibers (n = 28), of which eight studies were done in combination with probiotic supplementation. The most used interventions were fibers containing inulin-type fructans (i.e., inulin and fructooligosaccharides) followed by beta-glucan. Duration of interventions ranged from 4-52 weeks. Body weight reduction was associated with statistically significant changes in 54% of studies (n = 33). Body composition was reported in 41 studies and found fiber supplementation was associated with changes in waist circumference (n = 19 studies), visceral fat (n = 6), fat mass (n = 16), and lean mass (n = 5). Statistically significant changes in satiety measures (n = 7), glycemia and insulin parameters (n = 14), and microbial communities (n = 8) were also reported.

Conclusion: The results of this systematic review support the use of fiber supplementation as an adjunctive strategy to support weight management in adults with overweight/obesity. Interventions aimed at increasing fiber intake for adults with elevated BMI could result in improved body composition and health benefits.

Radha Chada, PhD, RD¹; Jaini Paresh Gala, MS²; Ashwini Chandrasekaran, MSc¹; Monish Karunakaran, MS, DrNB¹; G V Rao, MS, MAMS, FRCS³; Pradeep Rebala, MS¹; Balakrishna Nagalla, PhD⁴

¹Asian Institute of Gastroenterology, Hyderabad, Telangana; ²Asian Institute of Gastroenterology, Hyderabad, Telangana; ³Asian Institute of Gastroenterologist, Hyderabad, Telangana; ⁴Apollo Hospitals Educational and Research Foundation, Hyderabad, Telangana

Financial Support: None Reported.

Background: Sarcopenia is consistently recognized as a prognostic factor in chronic diseases and is linked to increased mortality in cancer patients due to both reduced muscle mass and function. This study aimed to assess the diagnosis of sarcopenia and its implications on post-operative outcomes among patients undergoing pancreaticoduodenectomy for pancreatic cancer.

Methods: Total skeletal muscle area (SMA), total skeletal muscle index (SMI) derived from abdominal computed tomography (CT) scans, and handgrip strength (HGS) were quantified using sex-specific Asian sarcopenia criteria. The impact of sarcopenia and the 6-minute walk distance (6-MWD) on hospital length of stay (LOS) and mortality post-surgery were analyzed.

Results: A total of 122 patients (91 males, 31 females) with an average age of 57.3 ± 10.79 years underwent assessment. The prevalence of low SMI across all age groups was 42.6%, with higher rates in females (71%) compared to males (33%). Decreasing BMI correlated with increasing prevalence of low SMI: 84.6% in underweight, 53.8% in normal weight, and 28.6% in overweight and obese patients (p < 0.00). Low HGS was more prevalent in females (48.4%) than males (29.7%), largely due to lower SMI. Both low SMI (p = 0.021) and low HGS (p < 0.00) increased with age. The prevalence of sarcopenia (18.9%), defined as low HGS and SMI, increased with age from 10% in < 53 years and 54-61 years age groups to 34.9% in >62 years age group (p = 0.004). Females exhibited a significantly higher prevalence of sarcopenia (41.9%) compared to males (11%) (p < 0.00). At admission, 79% of patients had 6-MWD values below reference levels. Mortality was significantly correlated with sarcopenia, with a higher mortality rate of 57.1% in sarcopenic patients compared to 42.9% in non-sarcopenic patients (p = 0.008). However, LOS did not show a significant correlation with sarcopenia. Additionally, 6-MWD did not correlate with either mortality or LOS.

Conclusion: Routine assessment for sarcopenia and frailty risk should be integrated into standard patient care protocols. Early identification of sarcopenia and monitoring of 6-MWD could offer a therapeutic window where timely interventions can be beneficial.

Krista Haines, DO, MA¹; Carrie Dombeck, MA²; Shauna Howell, BSN³; Trevor Sytsma, BS⁴; Jeroen Molinger, PhDc⁵; Amy Corneli, PhD²; Kenneth Schmader, MD⁴; Paul Wischmeyer, MD, EDIC, FCCM, FASPEN⁶

¹Duke University School of Medicine, Durham, NC; ²Duke University, Population Health Sciences, Durham, NC; ³Duke University, Department of Trauma Critical Care and Acute Care Surgery, Durham, NC; ⁴Duke University, Durham, NC; ⁵Duke University Medical Center - Department of Anesthesiology - Duke Heart Center, Raleigh, NC; ⁶Duke University Medical School, Durham, NC

Financial Support: Support was funded by a grant through the NIA Duke Pepper Older Americans Independence Center (Duke OAIC) and an investigator-initiated grant through Abbot Nutrition.

Background: Malnutrition is prevalent in older adult trauma patients, increasing the risk of poor outcomes and complicating recovery. A significant portion of these patients face nutritional deficiencies upon hospital admission, so tailored nutritional interventions are crucial. The SeND Home program is designed to bridge this gap by delivering structured, personalized nutritional support aimed at improving post-discharge recovery outcomes.

Methods: This study was a pilot randomized controlled trial with a mixed-method design, involving 40 older trauma patients who were randomized into the SeND Home intervention or standard care control groups using a 3:1 randomization. The intervention group received individualized nutrition, guided by Indirect Calorimetry (IC) to optimize caloric and protein intake, along with oral nutrition supplements (ONS) for one month post-discharge. Qualitative data were gathered through in-depth interviews with 20 participants from the intervention group to assess their experiences with the nutritional intervention, including perceived benefits, barriers, and long-term use of nutritional shakes.

Results: The SeND Home program demonstrated high feasibility and acceptability, with strong protocol adherence and positive reception from patients and healthcare providers. Quantitative analysis revealed that patients in both the intervention and control groups successfully completed key functional assessments, including six-minute walk tests and sit-to-stand measures, both in-person and virtual follow-up assessments, demonstrating the feasibility of conducting these measures remotely. This adaptability supports the study's potential for broader application across different care settings. Qualitative analysis provided deeper insights into patient experiences. Most participants expressed satisfaction with the ease of incorporating nutritional shakes into their daily routines, noting the convenience and palatability of the supplements. Several participants commented that the shakes helped them maintain their caloric and protein intake when regular meals were challenging. A recurring theme was the recognition of the importance of protein in recovery, with many participants stating that the shakes helped them meet their nutritional goals. However, some participants identified barriers to long-term use, including concerns about gastrointestinal discomfort, caloric content, and the cost of continuing supplementation after the study. Nonetheless, a majority reported they would consider using the shakes beyond the intervention period due to the perceived health benefits.

Conclusion: The SeND Home program offers a promising approach to addressing the nutritional needs of older adult trauma patients, with the potential to improve recovery outcomes through personalized nutritional support. The combination of quantitative and qualitative findings underscores the value of individualized nutrition, while qualitative insights emphasize the need to address patient-specific barriers to optimize long-term adherence and efficacy. Future studies with larger sample sizes are warranted to confirm these findings and further refine the SeND Home protocol to enhance patient-centered care in this vulnerable population.

Resilience manifests as a fluid reaction to acute or chronic health challenges.1 What we term "pre-stress reserve" comprises the characteristics a person possesses before facing a stressor, shaping their reaction to it. This capacity encompasses diverse dimensions, including psychological, physiological, and cognitive faculties leveraged to adapt positively to health-related stressors. Post-stressor reactions can be monitored through fluctuating metrics in multiple functional or health domains, such as cognitive capabilities, mobility, or emotional state, over a given timeframe. This entire sequence of interactions unfolds against the backdrop of external environmental influences. Enhanced resilience in confronting stressors is anticipated to yield more favorable long-term outcomes. Asterisk symbolizes potential intervention points at various stages, before, during, or after the exposure to a stressor, trauma.

Figure 1. Conceptual Model for Structured Nutrition Delivery Pathway Intervention.

Best of ASPEN-Malnutrition and Nutrition Assessment

Abstract of Distinction

Manpreet Mundi, MD¹; Osman Mohamed Elfadil, MBBS¹; Danelle Johnson, MS, RDN¹; Christopher Schafer, MS, RDN¹; Jami Theiler, RDN¹; Jason Ewoldt, MS, RDN¹; Madelynn Strong, MS, RDN¹; Katherine Zeratsky, RDN¹; Angie Clinton, MS, RDN¹; Sara Wolf, RDN¹; Ryan Hurt, MD, PhD¹

¹Mayo Clinic, Rochester, MN

Financial Support: None Reported.

Background: Creating personalized meal plans requires a deep understanding of an individual's nutritional needs, lifestyle, preferences, and health. Registered dietitian nutritionists traditionally handle this task, tailoring dietary recommendations for each patient. However, this is time-consuming, with data suggesting that outpatient dietitians spend an average of about an hour per patient. Artificial Intelligence (AI) has the potential to revolutionize many fields, including nutritional science, due to its ability to analyze large amounts of data and make predictions. Large language models (LLM), a type of AI, are trained on extensive textual datasets, allowing them to generate appropriate contextual text. This utility study aims to examine the utility of LLM in creating meal plans based on specific clinical questions compared to dietitians.

Methods: The study compared meal plans created by five clinical dietitians and four LLMs [Gemini (Google), CoPilot (Microsoft), ChatGPT 4.0 (Open AI), and a customized Chat GPT 4-0 that utilized specific sources for its data]. Five clinical scenarios were developed and validated for relevance. The dietitians and LLMs were asked to create 3-day meal plans based on those scenarios (Table 1). The dietitians recorded the time required to develop the meal plan. The dietitians and LLMs were asked to rate their comfort level on a scale of 1-5, with 5 being very comfortable. Three independent dietitians who were blinded to the author of the meal plan then utilized Nutritionist Pro to capture the macronutrient content of the meal plans. Accuracy in macronutrient goals and time taken to develop the meal plans were analyzed to examine differences between LLM- and RDN-generated meal plans.

Results: All LLMs and RDNs were able to provide a meal plan for each clinical scenario provided (Table 1). RDNs were generally comfortable generating the meal plans, with three feeling neutral to comfortable (mean comfort level score ranging between 3.8 and 4.2), while 2 rated their comfort level as comfortable to very comfortable (mean comfort level score ranging between 4.4 and 4.8). Overall, on average, the dietitians took the longest time to provide a meal plan for case (3) (61.8 ± 31.3 minutes), followed by case (1) (56.4 ± 52.7 minutes) (Table 2). LLMs developed a meal plan in less than 1 minute in all cases. In an analysis of variance, the RDN group had significantly higher accuracy in meeting the caloric targets for meal plans developed for all clinical scenarios (p-value < 0.001; Table 3). Notably, there was no difference between AI and RDN groups in the accuracy of daily calories meeting the specified goal for case 2 involving a Mediterranean diet (94.2% ± 13.2 In the AI group vs. 94.8% ± 15 in the RDN group; p-value = 0.895). There was no overall difference between AI and RDN groups in the accuracy of protein targets in meal plans (p-value = 0.215) (Table 3). Notably, case 4, which required lower protein in the setting of chronic kidney disease, was the only scenario in which AI LLMs provided significantly lower protein compared to the meal plan developed by the RDN group (Table 3). No difference was noted in the accuracy of AI and RDN in providing low carbohydrate, fat, or sodium diets when pertinent in clinical scenarios (Table 3).

Conclusion: Our findings suggest that examined LLMs are quite efficient in generating 3-day meal plans under various clinical scenarios. However, they fell short of meeting calorie goals in most scenarios and protein goals in one scenario. Based on current findings, the best approach may be to utilize LLMs to generate an initial meal plan with verification and adjustments made by a dietitian to ensure accuracy. More research is needed to explore AI applications in clinical nutrition.

Table 1. Clinical Scenarios.

Table 2. Time and Level of Comfort/Confidence.

Table 3. Accurate Attainment of Macronutrient Goals Through AI- and RDN-Generated Meal Plans.

Best International Abstract

Yoko Sakamoto, MD, PhD¹

¹Osaka University, Suita, Osaka

Financial Support: JSPS KAKENHI (grant numbers: JP16H06950, JP17K17854, JP21K08050).

Background: Elevated resting energy expenditure (REE) promotes cachexia, worsening prognosis in patients with advanced heart failure (HF). However, adequate assessment of energy balance is challenging because of unvalidated common prediction methods and unestablished determinants of REE, resulting in a lack of biomarkers for predicting insufficient energy intake. The objective of this cross-sectional study is to evaluate REE in patients with advanced HF and explore biomarkers for insufficient energy intake.

Methods: We measured REE by indirect calorimetry and calculated the total energy expenditure (TEE) of 72 hospitalized patients with advanced-stage HF. We compared these values with commonly-used formulas and analyzed the associations between REE per body weight (REEBW) and parameters related to hemodynamics and HF severity. In 17 of 72 patients, plasma amino acid concentrations and 24-hour urinary amino acids excretions were measured to analyze their correlations with energy balance, the ratio of caloric intake to REE.

Results: Patient Characteristics are summarized in Table 1. This study primarily included patients with advanced HF, as evident in 47 (65%) patients on the transplant waiting list and 61 (85%) patients with stage D HF. REE and TEE values were significantly higher than the predicted values. The mean REEBW was 25 kcal/kg/day, while that for the underweight (<18.5 kg/m²) was 28 kcal/kg/day. We found a significant negative correlation between REEBW and body mass index (BMI) (Figure 1), but no significant correlation between REEBW and HF-related parameters. The difference between TEE and predicted TEE using the European Society for Clinical Nutrition and Metabolism formula was most significant in the underweight patients because of underestimation, whereas TEE and predicted TEE using our modified formula with coefficients by BMI categories did not differ (Figure 2). There was a significant correlation between energy balance and urinary histidine and its metabolite 3-methylhistidine excretion, but no significant correlation with serum albumin and other amino acids concentrations (Figure 3).

Conclusion: Underweight patients with advanced HF require more energy per weight than the predicted value. Our proposed formula for predicted TEE in each BMI category may be useful in clinical practice to avoid underestimation of daily energy requirements. Inadequate energy intake, even with such an approach, may be identified by decreased urinary essential amino acids levels.

Table 1. Clinical Characteristics of Patients With Chronic Failure.

*p < 0.01, r: Pearson's or Spearman's coefficient value, p value in b) by unpaired t test.

Figure 1. Relationships Between Resting Energy Expenditure Per Body Weight (REEBW) and A) Age, B) Sex, and C) Body Mass Index (BMI).

The error bar represents SD **p < 0.01 by Tukey's honestly significance difference test.

Figure 2. Comparison of A) resting energy expenditure per body weight (REEBW), B) the difference between TEE and pTEE by the European Society for Clinical Nutrition and Metabolism (ESPEN) formula, C) the difference between total energy expenditure (TEE) and predicted value of TEE (pTEE) by the Harris-Benedict (HB) equation, D) the difference between TEE and pTEE by the Mifflin-St Jeor equation, and E) the difference between TEE and pTEE (proposed formula), among each group divided by body mass index.

*p < 0.05, r: coefficient value by Pearson's correlation analysis, p value by Pearson's correlation test.

Figure 3. Relationships between the ratio of energy intake to resting energy expenditure (REE) and the serum levels of A) albumin, B) essential amino acids, C) histidine, D) 3-methylhistidine, E) threonine, and F) lysine, and urinary levels of G) histidine, H) 3-methylhistidine, I) threonine, and J) lysine.

Trainee Award

Abstract of Distinction

Raheema Damani¹; Shubha Vasisht¹; Valerie Luks, MD¹; Yue Ren, PhD¹; James Rowe¹; Charlene Compher, PhD¹; Jeffrey Duda, PhD¹; James Gee, PhD¹; Rachel Kelz, MD¹; Hongzhe Li, PhD¹; Gary Wu, MD¹; Walter Witschey, PhD¹; Victoria Gershuni, MD¹

¹University of Pennsylvania, Philadelphia, PA

Financial Support: None Reported.

Background: Malnutrition is associated with poor outcomes after abdominal surgery and increased perioperative morbidity. Despite recognition that pre-operative nutrition interventions can improve outcomes, current practice does not routinely screen for or assess malnutrition pre-operatively, hence missing an opportunity to intervene. This study performed machine learning-based quantitative assessment of clinically obtained pre-operative abdominal computed tomography (CT) scans to develop novel imaging-derived phenotypes (IDPs) of muscle and fat volumes. We further developed predictive models for pre-operative screening of clinical malnutrition among patients undergoing abdominal surgery.

Methods: A retrospective analysis (2018-2021) of patients undergoing abdominal surgery at a single quaternary care institution was conducted. Outcomes collected for participation in American College of Surgery National Surgical Quality Improvement Program were augmented with nutritional assessment variables and pre-operative abdominal CT scan. Imaging features were derived using a novel machine-learning algorithm, to automate quantification of size (height-adjusted volume) and attenuation (HU) of five muscle groups and two fat depots. Body composition features of sarcopenia, myosteatosis, and visceral obesity were determined based on predefined cutoffs from literature. Logistic regression identified features associated with nutrition status. Sex-specific elastic net regression models were developed to predict diagnosis of clinical malnutrition using a combination of clinical and imaging features. Model performance was evaluated using the area under the receiver operating curve (AUROC). The DeLong test assessed significance between models. Clinical utility was assessed via decision curve analysis.

Results: Out of 1143 patients, 20.2% (n = 231) were diagnosed with clinical malnutrition. Patients with clinical malnutrition had increased odds of post-operative complication (OR = 2.7, p < 0.001) and prolonged length of stay (OR = 4.5, p < 0.001). CT revealed high prevalence of sarcopenia (55.8%), myosteatosis (54.5%) and visceral obesity (63.3%). Males and females had distinct differences in body composition. Decreased muscle size was associated with malnutrition in both sexes, but only females had significant differences in muscle quality (attenuation). Adjusted logistic regression demonstrated that multiple imaging features were associated with increased odds of malnutrition in a sex-specific manner (Figure 1B). Using elastic net regression to determine the likelihood of clinical malnutrition, a multimodal model that incorporated imaging features outperformed a model with clinical features alone (males: AUC: 0.76 vs. 0.79, p < 0.05, females: 0.70 vs. 0.78, p < 0.05). Decision curve analysis revealed higher net benefit for the multimodal model, indicating clinical utility for pre-operative imaging-based malnutrition screening.

Conclusion: Machine learning-based quantitative assessment of pre-operative CT scans can be utilized to develop models for the pre-operative period to screen for clinical malnutrition. Implementation of CT-based automated pre-operative nutrition screening will create a window of opportunity for more targeted peri-operative nutrition intervention to reduce the risk of post-operative adverse outcomes.

Critical Care and Critical Health Issues

Trainee Award

International Abstract of Distinction

Tomonori Narita, MD¹; Kazuhiko Fukatsu, MD, PhD²; Satoshi Murakoshi, MD, PhD³; MIdori Noguchi, BA⁴; Reo Inoue, MD, PhD⁵; Nana Matsumoto, RD, MS⁵; Seiko Tsuihiji, BA⁵; Toshifumi Asada, MD, PhD⁵; Miyuki Yamamoto, MD, PhD⁵; Ryohei Horie, MD, PhD⁵; Ryota Inokuchi, MD, PhD⁵; Shoh Yajima, MD, PhD⁵; Koichi Yagi, MD, PhD⁵; Kento Doi, MD, PhD⁵; Yoshifumi Baba, MD, PhD⁵

¹The University of Tokyo, Chuo-City, Tokyo; ²The University of Tokyo, Bunkyo-ku, Tokyo; ³The Kanagawa University of Human Services, Yokosuka-City, Kanagawa; ⁴The University of Tokyo Hospital, Bunkyo-ku, Tokyo; ⁵The University of Tokyo, Bunkyo-City, Tokyo

Financial Support: None Reported.

Background: Early enteral nutrition (EN) is recommended for critically ill or severely injured patients. However, feeding the ischemic gastrointestinal tract with EN may lead to gut necrosis. Although various symptoms have been used to assess gut tolerance to EN, none of them provide conclusive evidence of gut ischemia. Here, we conducted a prospective study to investigate the blood flow (BF) in the superior mesenteric artery/vein (SMA/SMV) of patients in intensive care unit (ICU) using ultrasonography along with traditional physical findings used to assess gut function.

Methods: We enrolled 60 patients who were admitted to the ICU at The University of Tokyo Hospital from July 2023 to June 2024 (approved by the Ethics Committee of the University of Tokyo under protocol No. 2023049NI). SMA and SMV BF, physical findings, hematological findings and course after admission were evaluated within 24 and 48 hours after admission and on the 4th and 7th days after admission. SOFA and APACHE II scores were also measured. The kinetics of this BF was compared between patients with and without EN intolerance or between patients who survived 10 days after admission to ICU and not. The relationship between the flow and other parameters were analyzed using linear mixed models and regression analysis.

Results: EN intolerance only occurred in two cases. There were no significant differences in SMA and SMV BF between patients with EN intolerance and those without intolerance. And none of the gut necrosis was observed. There were no significant correlations between other physical/hematological findings and SMA/V BF (linear mixed model and regression analysis, Table 1) at any time point. Seven patients died within 10 days after admission, and in these patients, SMV BF and the SMV/SMA BF ratio within 24 hours after admission were significantly reduced compared to patients who survived (SMV: p = 0.0044, SMV/SMA: p = 0.0089, linear mixed model, Figure 1). Additionally, reduced SMV BF continuing until 48 hours was associated with early death (p = 0.0285). Receiver operating characteristic curve (ROC) analysis with early death as the endpoint indicated that SMV BF and the SMV/SMA BF ratio within 24 hours after admission had higher area under the curve (AUC) (SMV: AUC = 0.83766, SMV/SMA: AUC = 0.81063) than SOFA and APACHE II score (SOFA: AUC = 0.50539, APACHE II: AUC = 0.54043).

Conclusion: Because physical findings or hematological findings did not correlate with gut BF, routine measurement of SMA/SMV BF may be proposed to prevent gut necrosis resulting from EN under gut ischemic condition. Patients who died early after admission had significantly lower SMV BF and SMV/SMA BF ratio at admission than survivors. AUC of SMV BF and SMV/SMA ratio were higher than existing prognostic indicators such as the SOFA and APACHE II scores, suggesting that SMA and SMV BF measurements may be usable to assess the risk mortality after admission.

Table 1. Correlation Between Physical/Hematological Findings and SMA/V BF.

Figure 1. The Temporal Changes of SMV BF and SMA/SMV BF Ratio After Admission.

International Abstract of Distinction

Fengchan Xi¹; Nan Zheng¹; Bing Xiong²; Di Wang¹; Teng Ran¹; Xinxing Zhang¹; Tongtong Zhang¹; Caiyun Wei¹; Xiling Wang³; Shanjun Tan⁴

¹Department of Intensive Care Unit, Women's Hospital of Nanjing Medical University (Nanjing Women and Children's Healthcare Hospital), Nanjing, Jiangsu, China; ²Department of Radiology, The First Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, Zhejiang Province, China; ³Department of Radiology, The First Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, Zhejiang Province, China, Shanghai, Jiangsu; ⁴Department of General Surgery/Shanghai Clinical Nutrition Research Center, Zhongshan Hospital, Fudan University, Shanghai, China

Financial Support: None Reported.

Background: Skeletal muscle density (SMD) is a valuable prognostic indicator in various conditions such as cancer, liver cirrhosis. Yet, the connection between SMD and intra-abdominal infection in individuals who have suffered abdominal injuries is still unclear. The purpose of this research is to examine how well SMD can predict intra-abdominal infection in patients who have suffered abdominal trauma.

Methods: Participants with abdominal injuries were included in this research from January 2015 to April 2023. Based on the quartile of SMD, the entire population was split into two categories (Figure 1). Prognostic factors were identified through logistic regression analysis. ROC was used to assess the predictive accuracy of SMD and its combinations with other biomarkers for clinical outcomes.

Results: A total of 220 patients were ultimately included in the study (Figure 2). Patients in the group with low SMD exhibited a higher incidence of intra-abdominal infection, longer hospital stays, and increased hospital costs (Table 1). In patients with abdominal trauma, low SMD was identified as a significant independent predictor of intra-abdominal infection (OR 2.510; 95% CI 1.168-5.396, p = 0.018). Low SMD had a higher area under the curve (AUC) in ROC analysis compared to TRF, NRS2002 score, and APACHEII score for predicting intra-abdominal infection (AUC 0.70, 95% CI 0.61-0.78, p = 0.002). Moreover, low SMD showed associations with clinical outcomes such as hospital stay length and costs (Figure 3, p < 0.01).

Conclusion: Low SMD is recognized as an independent risk factor for predicting intra-abdominal infections in this patient population. Notably, SMD is emerging as a novel predictor of abdominal infections in patients with abdominal trauma.

Table 1. Demographic Information and Clinical Characteristics of the Study Population.

Data were expressed as number (percentage) of patients, mean ± standard deviation or median and quartile range. BMI, Body Mass Index; SMD, skeletal muscle density; HU, Hounsfield Unit; SMI, skeletal muscle mass index; SMA, skeletal muscle areas; PCT, procalcitonin; CRP, C-reactive protein; TRF, transferrin; RBD, retinol binding protein; ALB, albumin; DIC, Disseminated intravascular coagulation; GCS, Glasgow coma scale; ISS, Injury severity score; NRS, Nutritional Risk Screening; SGA, Subjective Global Assessment; APACHE, Acute physiology and chronic health evaluation; SOFA, Sequential organ failure assessment; ICU, Intensive Care Unit.

A man 65 years old with SMD 46.61 Hounsfield Unit (HU) (A-C) and a man 31 years old with SMD 20.88 (HU) (D-F) underwent measurements of muscle and fat distribution at the level of the third lumbar spine (L3) using computed tomography (CT). Muscle tissue was highlighted in red.

Figure 1. Computed Tomography (CT) Measurement of Skeletal Muscle Density (SMD) at the Third Lumbar Vertebra (L3).

Figure 2. Study Flow Chart.

SMD, skeletal muscle density; TRF, transferrin; NRS, Nutritional Risk Screening; APACHE, Acute physiology and chronic health evaluation; LOS, Length of Stay, **p ＜ 0.01, *p＜ 0.05.

Figure 3. Correlation Analysis fo SMD, LOS and Hospital Cost.

Lucia Gonzalez Ramirez, MCN¹; Jessica Alvarez, RD, PhD¹; Dean Jones, PhD¹; Thomas Ziegler, MD²

¹Emory University, Atlanta, GA; ²Emory Healthcare, Atlanta, GA

Financial Support: None Reported.

Background: Mortality rates for intensive care unit (ICU) patients in the United States range from 10-30%. A better understanding of the impact of critical illness on metabolism (e.g., metabolites and related metabolic pathways) associated with ICU mortality is needed. Our aim in this pilot study was to evaluate the global metabolome in critically ill adults requiring surgical ICU (SICU) care and parenteral nutrition (PN) using plasma high-resolution metabolomics (HRM).

Methods: Secondary analysis of a completed, pragmatic, randomized, controlled, multicenter trial of glutamine-supplemented vs. standard glutamine-free PN, conducted in adults requiring SICU care after cardiac, vascular, or gastrointestinal surgery. Plasma HRM from 11 participants who died within 28 days after study enrollment was compared to 13 participants who survived over the same timeframe. Plasma was obtained at baseline, day three, and day seven after study enrollment and analyzed using liquid chromatography coupled with high-resolution mass spectrometry in C18^- electrospray mode. A metabolome-wide association study (MWAS) was conducted using multiple linear regression models to assess the relationship between the plasma metabolome and survivorship status. Pathway enrichment analysis and a meet-in-the-middle approach were performed to characterize metabolic pathways and intermediate metabolic features (metabolites) between groups.

Results: Among 7,146 plasma metabolic features used in downstream analysis, 184 exhibited a significant change (raw p-value < 0.05) from baseline (study entry) to day three and were linked to survivorship status. Similarly, 282 metabolic features were selected from baseline to day seven and linked to survivorship status. From these metabolomic features, the meet-in-the-middle analytic approach identified 31 overlapping metabolites linked to five known human metabolic pathways, including fatty acid activation and oxidation and de novo fatty acid biosynthesis. In addition, pathway enrichment analysis showed that C21-steroid hormone biosynthesis and metabolism, branched-chain amino acid degradation, metabolism of threonine, methionine, cysteine, and other amino acids, and vitamin E-related metabolic pathways were significantly linked to survivorship status, Figure 1.

Conclusion: Using plasma HRM and MWAS, with complimentary pathway enrichment analysis, we identified steroid hormone, vitamin E, and numerous macronutrient-related metabolic pathways significantly associated with survivorship in SICU patients requiring PN. Larger prospective studies are needed to confirm these results and to identify specific pathway-related metabolites associated with ICU survival in patients requiring PN. In addition, well-powered studies to determine the impact of altered PN amino acid composition on systemic metabolism are needed. Plasma HRM may be a useful tool to understand nutrition-related pathophysiology in critically ill patients requiring specialized feeding.

Figure 1.

Vishal Chandel, MD¹; Kris Mogensen, MS, RD-AP, LDN, CNSC²; Marielle Austen, RD, LDN, CNSC²; Diane Herzog, MS, RD-AP, LDN, CNSC²; Malcolm Robinson, MD³

¹Brigham and Women's Hospital, Harvard Medical School, Boston, MA; ²Brigham and Women's Hospital, Boston, MA; ³Brigham and Women's Hospital, Harvard Medical School, Boston, MA

Financial Support: None Reported.

Background: In decompensated liver disease, hyperammonemia is considered the underlying metabolic derangement. However, non-cirrhotic hyperammonemia (NCH) is less common. Urea cycle disorders (UCD) and infections by urease producing organisms should be considered for unexplained hyperammonemia. We present a case of NCH complicated by cerebral edema with failure conventional therapy. A 45-year-old female with multiple sclerosis (on monoclonal antibody), neuromyopathy and seronegative inflammatory arthritis presented with acute abdominal pain. Imaging revealed pneumoperitoneum from gastric perforation secondary to chronic steroid use. The course was further complicated by intraabdominal sepsis due to gastric leak. During treatment, patient developed progressively elevated ammonia levels of unclear etiology as the patient had no history of liver disease. Ammonia peaked at 676 μmol/L (normal range 11–60). Labs and imaging demonstrated no cirrhosis or acute liver failure. Patient was started on lactulose and then to continuous renal replacement therapy for ammonia clearance. Ammonia remained elevated despite these treatments. Encephalopathy ensued with neurologic deterioration requiring intubation. The patient required parenteral nutrition (PN) due to gastric leak and recent wedge resection with need to hold enteral nutrition. PN initially provided amino acids at 1.5 g/kg, patient received total seven days of PN before her ammonia level started trending upwards. Amino acids were removed from the PN when the ammonia level was 326 μmol/L. Patient was found to have positive rare Ureaplasma species, later confirmed as Ureaplasma Parvum. Given her deteriorating condition and lack of susceptibility testing in real-time; azithromycin, doxycycline and levofloxacin were started. However, patient continued to decline with immunocompromised status, bowel perforation and progressive mental status worsening eventually leading to cerebral edema, herniation and death. Bacterial degradation of glutamine in small intestine and of protein and urea in colon are major ammonia sources in our body. Elevated ammonia levels occur in UCDs, portosystemic-shunting, urinary tract infection from urease-producing organisms, ureterosigmoidostomy, shock, renal disease, heavy exercise, smoking, gastrointestinal bleeding, salicylate intoxication, medications like valproate and 5-Fluorouracil. In our case, patient had NCH non-responsive to treatment. The mechanism is likely related to having an undiagnosed inborn error of metabolism (IEM) unmasked by infection with urea producing organism in the setting of severe immunocompromised status and post-surgical complications. In this patient, chronic steroid and immunomodulator use unmasked IEM by providing a flourishing medium for urease-producing organisms. Although IEMs often have early age of onset, UCDs have multiple modes of inheritance and can present at later age. Ornithine transcarbamylase deficiency is the most common UCD and can present with reduced plasma citrulline in both children and adults. Treating a potential IEM begins prior to confirmation of an etiology. Treatment should start early if UCD is suspected. Early treatment includes holding protein delivery and providing dextrose and lipids only, with energy delivery targeted at 35-40 kcal/kg to avoid muscle catabolism. Prompt treatment with ammonia scavengers, renal replacement to reduce ammonia levels, replacing urea cycle substrates and reducing catabolic state are essential. Additional nutritional therapies, such as treatment with essential amino acids may be required. Timely intervention is essential to prevent mortality.

Methods: None Reported.

Results: None Reported.

Conclusion: None Reported.

Table 1. Patient's Trend of Ammonia Levels During Her Course of Hospitalization, With Parental Nutrition Being Stopped on Day 3 of Ammonia Rise.

(CPS-1, OTC, ASS, ASL, and ARG are enzymes). CPS-1, carbamoyl phosphate synthetase-1; OTC, ornithine transcarbamylase; ASS, argininosuccinate synthetase; ASL, argininosuccinate lyase; ARG, arginase.

Figure 1. Ammonia Metabolism in Urea Cycle.

Image Credit: VectorMine/Shutterstock.com.

Figure 2. Urea Cycle Pathway in Human Body.

Mateen Jangda, BS, MS¹; Hannah Kittrell, RD¹; Jaskirat Gill, MD¹; Ahmed Shaikh, MD¹; Rebecca Wig, MD²; Rohit Gupta, MD³; Shruti Bakare¹; Roopa Kohli-Seth, MD⁴; Paul McCarthy, MD⁵; Jayshil Patel, MD⁶; Girish Nadkarni, MD⁴; Ankit Sakhuja, MBBS, MS⁴

¹Mount Sinai, New York, NY; ²U Arizona, Tucson, AZ; ³Mount Sinai, New York, NY; ⁴Icahn School of Medicine at Mount Sinai, New York, NY; ⁵West Virginia University, Morgantown, WV; ⁶Medical College of Wisconsin, Milwaukee, WI

Financial Support: NIH/NIDDK K08DK131286 - provided to Ankit Sakhuja.

Background: The provision of enteral nutrition (EN) is a key component of managing mechanically ventilated patients in the intensive care unit (ICU). Despite critical care nutrition guideline recommendations to achieve at least 70% of daily caloric requirements (eucaloric) within 3-7 days of critical illness, many critically ill, mechanically ventilated patients do not achieve this goal. We aimed to assess the feasibility of a deep learning model to identify and predict, starting from day 3 of intubation, the likelihood that critically ill, mechanically ventilated patients will fail to achieve at least 70% of daily caloric requirements through EN.

Methods: In this retrospective study, using the MIMIC-IV database, we identified adult ICU patients ( ≥ 18 years old) who were mechanically ventilated for at least 72 hours and received EN. We excluded patients that received parenteral nutrition. As per ASPEN guidelines, the daily caloric requirement was defined as 11-14 kcal/kg actual body weight/day for BMI of 30-40kg/m2, 22-25 kcal/kg ideal body weight/day for BMI > 50kg/m2, and 25 kcal/kg actual body weight/day for the remaining patients. Receipt of eucaloric nutrition was defined by achieving at least 70% of daily caloric requirements from both EN and propofol, a sedative contained in a fat emulsion that also provides calories. To develop our deep learning model, we utilized readily available electronic health record data, including demographics, comorbidities, vital signs, administered medications (eg, vasopressors, intravenous fluids, sedatives, and pain medications), net fluid balance, and EN data. We divided the cohort into an 80% training set, 10% validation set, and a 10% hold-out test set. We developed a multi-input, multi-output Long Short-Term Memory (LSTM) network to predict the likelihood of a patient failing to achieve eucaloric nutrition for each day starting with day 3 after intubation and continuing until day 7, ICU discharge, or extubation, whichever occurs first. Predictions were made every 4 hours. The model was trained on the training cohort, optimized using the validation cohort by fine-tuning model parameters and hyperparameters during training, and validated on the hold-out test set.

Results: The study cohort included 5,097 mechanically ventilated ICU patients, divided into training (80%, n = 4,077), validation (10%, n = 510), and test (10%, n = 510) sets. The mean age was 63.75 years, with 58.39% being male and 60.68% identifying as white. More than one-third of patients did not meet eucaloric nutrition goals at any time point. The percentage of patients not achieving eucaloric enteral nutrition each day are shown in Table 1. The LSTM model showed strong predictive performance on the hold-out test set, achieving a Receiver Operating Characteristic (ROC) Area Under the Curve (AUC) of 0.8719 across all time steps, with an overall accuracy of 83.19%. ROC AUC for the model at each time-interval is shown in Figure 1.

Conclusion: Our study demonstrates the feasibility and accuracy of deep learning models to predict which mechanically ventilated ICU patients fail to achieve eucaloric nutrition from EN starting from day 3 after intubation. By identifying individuals at risk of not meeting their nutritional goals, our model can help facilitate timely and targeted nutritional interventions, improving the management of nutrition support in these patients. However, external validation is needed to confirm the model's generalizability, and further research is warranted to explore its integration into clinical workflows and assess its impact on patient recovery and outcomes.

Table 1. Percentage of Patients Not Achieving Eucaloric Nutrition Starting With Day 3 of Intubation.

Figure 1. ROC AUC of the LSTM Model at Each 4-Hour Time Interval of Prediction.

Abstract of Distinction

Jeroen Molinger, PhDc¹; Ibtehaj Naqv, PhD, MD²; Christina Barkauskas³; Krista Haines, DO, MA⁴; Marat Fudim⁵; David MacLeod⁶; John Whittle⁷; Henrik Endeman⁸; Manesh Patel⁵; Jan Bakker⁸; Paul Wischmeyer, MD, EDIC, FCCM, FASPEN⁹; Zachary Healy³

¹Duke University Medical Center - Department of Anesthesiology - Duke Heart Center, Raleigh, NC; ²Duke University Hospital, Durham, NC; ³Duke University Hospital, Dep. of Pulmonology, Durham, NC; ⁴Duke University School of Medicine, Durham, NC; ⁵Duke University Medical Center - Duke Heart Center, Durham, NC; ⁶Duke University Medical Center - Department of Anesthesiology - Duke Human Pharmacology and Physiology Lab (HPPL), Durham, NC; ⁷UCLH, London, England; ⁸Erasmus Medical Center University, Rotterdam, Zuid-Holland; ⁹Duke University Medical School, Durham, NC

Financial Support: None Reported.

Background: Hyperinflammation in critical illnesses like COVID-19 can lead to toll-like receptor (TLR) tolerance in monocytes, resulting in diminished immune responses and metabolic adaptability. This study explores the potential of exogenous ketone supplementation (EKS) to address monocyte metabolic dysfunction during critical illness. Ketone bodies, particularly β-hydroxybutyrate (D-BHB), serve as alternative energy substrates during low carbohydrate availability. EKS has shown promise in enhancing insulin sensitivity, reducing oxidative stress, and dampening inflammation by downregulating inflammatory pathways such as the NLRP3 inflammasome. This research investigates the effects of EKS on monocyte functionality in critically ill patients.

Methods: The study utilized a Seahorse-based mitochondrial stress test to evaluate monocytes' ability to respond to mitochondrial stress and glycolytic transition. Monocytes from healthy volunteers were compared to those from COVID-19 ICU patients. The cells were exposed to either no stimulus (control), a TLR-4 agonist (LPS), or a TLR-7 agonist (R848). Oxygen consumption (OCR) and extracellular acidification rate (ECAR) were measured to assess metabolic responses. Additionally, the effect of D-BHB pre-incubation on monocyte metabolic function was evaluated.

Results: Two-way ANOVA showed a significant difference “*” between all treatment groups at each time point when comparing healthy monocytes to the monocytes from a COVID-ICU survivor. Two-way ANOVA showed a significant difference between all treatment groups at all time points when comparing healthy monocytes to those from a COVID-ICU non-survivor (Figures 1 and 2). Follow-up unpaired t-tests showed a difference between healthy and survivor monocytes in each phase. However, a significant difference was only observed between healthy and non-survivors in the maximal respiration phase via unpaired t-test.

Conclusion: COVID-19 patient monocytes exhibited significantly impaired metabolic responses compared to healthy controls. Dramatically decreased OCR in COVID-19 monocytes, regardless of treatment. Disturbed ability to shift metabolism to a more glycolytic response under mitochondrial stress was seen. This proof-of-concept study demonstrates the potential utility of ketone monoesters in addressing metabolic dysfunction in monocytes during critical illness. The findings suggest that EKS could significantly alleviate the metabolic and immune dysfunctions associated with conditions like COVID-19. Impaired metabolic flexibility in monocytes from critically ill patients combined with Improved glycolytic responses and metabolic adaptability with D-BHB pre-treatment. T cell metabolic function can be augmented with ketone supplementation. By addressing the metabolic dysfunction in immune cells, EKS may offer a novel approach to improving outcomes for critically ill patients. Further investigation is warranted to fully elucidate this promising therapeutic strategy's mechanisms and potential benefits.

Figure 1. Monocyte Mitochondrial Stress From COVID-19 ICU Patients (Survivor and Non-Survivor) Compared to Healthy Volunteers.

Figure 2. OCR Data from Normal Monocytes Shows that a Ketone Monoester Can Nullify the Harmful Effects of R848.

GI, Obesity, Metabolic, and Other Nutrition Related Concepts

Vishal Chandel, MD¹; Kris Mogensen, MS, RD-AP, LDN, CNSC²; Patricia Laglenne, MS, RD, LDN²; Katherine McManus, MS, RD, LDN²; Malcolm Robinson, MD³

¹Brigham and Women's Hospital, Harvard Medical School, Boston, MA; ²Brigham and Women's Hospital, Boston, MA; ³Brigham and Women's Hospital, Harvard Medical School, Boston, MA

Financial Support: None Reported.

Background: Acute pancreatitis (AP) can lead to malnutrition due to hypercatabolism, alterations in glucose metabolism and imbalances in the water-electrolyte and acid-base status. Appropriate management & nutritional support can potentially affect outcome. Since inflammation plays a key role in initiation and progression of pancreatitis, nutrition therapy may modulate the oxidative stress response and counteract the catabolic effects. The optimal route of feeding remains a clinical question. This meta-analysis aims to better characterize enteral nutrition (EN) and parenteral nutrition (PN) in patients with pancreatitis and to assess the effect of nutrition on clinical outcomes.

Methods: We searched Medline, PubMed, Embase & Cochrane databases from 1965-2023 for relevant studies using terms: ‘enteral nutrition’, ‘tube feeding’, ‘artificial feeding’, ‘nasogastric’, ‘nasojejunal’, ‘parenteral nutrition’, and ‘acute pancreatitis’. Random effects model was used to conduct network meta-analysis for estimating risk ratios (RRs) with 95% confidence interval (CI) for primary outcomes of EN or PN in patients associated with: 1) infectious complications, 2) multi organ failure (MOF) and 3) mortality. RevMan 5.3 was used for meta-analysis; binary variables were statistically analyzed by Mantel-Haenszel method, using odds ratio (OR) and 95% confidence interval (CI) reported statistics, with P < 0.05 considered statistically significant.

Results: We identified 552 studies in the initial search. After applying exclusion criteria, 21 articles were included in the final quantitative analyses. From a total of 7,557 participants, 4,851 patients received PN and 2,708 patients received EN. Sixteen studies (7,003 participants) reported mortality outcomes. Patients receiving EN had lower odds of mortality compared to those receiving PN [odds ratio (OR) = 1.95,95% confidence interval (CI) 1.60-2.39) (p < 0.00001). Eighteen studies (n = 7,214 participants) reported infectious complications. EN was associated with a significantly lower incidence of infections (OR = 2.36, 95% CI: 1.36 to 4.08, p < 0.05). However, when 8 studies were excluded to lower the statistical heterogeneity (I² = 35%, p = 0.15), no significant difference was observed in the rate of infectious complications in PN vs. EN groups (OR = 1.09, 95% CI: 0.67 to 1.75, p > 0.05). Twelve studies (707 patients) reported MOF incidence rates, showing reduced MOF rates in EN group (OR = 3.76, 95% CI: 1.87 to 7.55, p < 0.0001). Twelve studies reported the hospital length of stay (LOS). Only 9 studies were included for final analysis to ensure improved statistical homogeneity (I² = 54%, p = 0.09) and showed shorter length of hospital stay in EN group than PN group (OR = 2.27, 95% CI: 1.45 to 4.09, p = 0.0001).

Conclusion: This meta-analysis shows that in patients with acute pancreatitis, EN is associated with improved outcomes including decreased incidence of MOF, decreased hospital LOS and decreased overall mortality compared to PN. These results must be interpreted with caution as they may be related to severity of the pancreatitis rather than the route of feeding. However, large sample sizes of these studies with low heterogeneity suggest that difference in severity of pancreatitis may only partially explain, if at all, the significant differences identified. Analyses of the results which formally control for severity of pancreatitis are warranted. These initial findings suggest that EN is the preferred route of nutrition support for the hospitalized patients with pancreatitis.

Table 1. Data Sheet on Demographic Information and Outcomes Data of the Included Studies in the Meta-Analysis Which Compared Parenteral Nutrition (PN) and Enteral Nutrition (EN) in Patients With Acute Pancreatitis.

Figure 1. Forest plot for sensitivity analysis of the effect of EN and PN on mortality in patients with AP, where PN is the experimental arm and EN is the control arm. EN, enteral nutrition; PN, parenteral nutrition; CI, confidence interval; AP, acute pancreatitis.

Figure 2. Forest plot for sensitivity analysis of the effect of EN and PN on the incidence of infectious complications in patients with AP, where PN is the experimental arm and EN is the control arm. EN, enteral nutrition; PN, parenteral nutrition; CI, confidence interval; AP, acute pancreatitis.

Figure 3. Forest plot for sensitivity analysis of the effect of EN and PN on the incidence of multiple organ failure in patients with AP, where PN is the experimental arm and EN is the control arm. EN, enteral nutrition; PN, parenteral nutrition; CI, confidence interval; AP, acute pancreatitis.

Figure 4. Forest plot for sensitivity analysis of the effect of EN and PN on the length of hospital stay in patients with AP, where PN is the experimental arm and EN is the control arm. EN, enteral nutrition; PN, parenteral nutrition; CI, confidence interval; AP, acute pancreatitis.

Salvador Ortiz-Gutiérrez, MSc, RD¹; Aurora Elizabeth Serralde-Zúñiga, MD, PhD²; Adriana Flores-López, PhD, RD²; Luis Eduardo González-Salazar, PhD, RD²; Maria Guadalupe Estrada-Trujillo, RD³; Edgar Pichardo-Ontiveros, MSc, RD³; Berenice Palacios-González, PhD, RD⁴; Héctor Infante-Sierra, MD, MSc⁵; Elena Juventina Tuna-Aguilar, MD⁶; Laura Alejandra Velázquez-Villegas, PhD, RD³; Andrea Ramírez-Coyotecatl, MD³; Sandra María Carrillo-Córdova, RD³; Karla Guadalupe Hernández-Gómez, PhD, RD²; Rocío Guizar-Heredia, MSc, RD³; Ana Vigil-Martínez, MSc, RD³; Isabel Medina, PhD, RD⁷; Azalia Ávila-Nava, PhD⁸; Angélica Borja-Magno, PhD, RD³; Juan Gerardo Reyes-García, MD, PhD⁹; Adriana Margarita López-Barradas, PhD, RD³; Andrea Díaz-Villaseñor, PhD¹⁰; Samuel Canizalez-Quinteros, PhD¹¹; Nimbe Torres, PhD³; Armando Roberto Tovar, PhD³; Martha Guevara-Cruz, MD, PhD³

¹Fisiología de la Nutrición (Nutrition Physiology), Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán (National Institute of Medical Sciences and Nutrition Salvador Zubirán), México; Sección de Estudios de Posgrado e Investigación (Postgraduate Studies and Research Section), Escuela Superior de Medicina (Higher School of Medicine), Instituto Politécnico Nacional (IPN) (National Polytechnic Institute), Distrito Federal (Mexico City), México; ²Nutriología Clínica (Clinical Nutrition), Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán (National Institute of Medical Sciences and Nutrition Salvador Zubirán), Distrito Federal (Mexico City), México; ³Fisiología de la Nutrición (Nutrition Physiology), Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán (National Institute of Medical Sciences and Nutrition Salvador Zubirán), Distrito Federal (Mexico City), México; ⁴Laboratorio de Envejecimiento Saludable (Laboratory of Healthy Aging), Instituto Nacional de Medicina Genómica (National Institute of Genomic Medicine), Centro de Investigación sobre Envejecimiento (Center for Research on Aging) (CIE-CINVESTAV), Distrito Federal (Mexico City), México; ⁵Hospital Central del Sur PEMEX (Hospital Central del Sur PEMEX), Distrito Federal (Mexico City), México; ⁶Hematología y Oncología (Hematology and Oncology), Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán (National Institute of Medical Sciences and Nutrition Salvador Zubirán), Distrito Federal (Mexico City), México; ⁷Metodología de la Investigación (Research Methodology), Instituto Nacional de Pediatría (National Institute of Pediatrics), Distrito Federal (Mexico City), México; ⁸Unidad de Investigación (Research Unit), Hospital Regional de Alta Especialidad Península de Yucatán IMSS-Bienestar (Yucatan Peninsula Regional High Specialty Hospital IMSS-Bienestar), Mérida México, Merida, Yucatan; ⁹Sección de Estudios de Posgrado e Investigación (Section of Postgraduate Studies and Research), Escuela Superior de Medicina (Higher School of Medicine), Instituto Politécnico Nacional (National Polytechnic Institute) (IPN), Distrito Federal (Mexico City), México; ¹⁰Medicina Genómica y Toxicología Ambiental (Genomic Medicine and Environmental Toxicology), Instituto de Investigaciones Biomédicas (Biomedical Research Institute), Universidad Nacional Autónoma de México (National Autonomous University of Mexico) (UNAM), Distrito Federal (Mexico City), México; ¹¹Unidad de Genómica de Poblaciones Aplicada a la Salud (Population Genomics Unit Applied to Health), Universidad Nacional Autónoma de México/Instituto Nacional de Medicina Genómica (National Autonomous University of Mexico/National Institute of Genomic Medicine), Distrito Federal (Mexico City), México

Financial Support: Gobierno de la Ciudad de México, Secretaría de Educación, Ciencia, Tecnología e Innovación (Government of Mexico City, Ministry of Education, Science, Technology and Innovation) (2150c23).

Background: Obesity is one of the major problems of public health worldwide, with several complications for those who live with this problem. Recent research has found that higher body mass index (BMI) is associated with lower serum iron concentrations. One of the main mechanisms explaining this relationship is low-grade chronic inflammation, linked with quantity and dysfunction of adipose tissue. Hepcidin is a peptide hormone that regulates the transit of iron in the organism. When hepcidin is increased, the iron-exporter ferroportin is blocked, resulting in iron trapped inside the cells and producing a deficiency state of iron because of inflammation. The existence of inflammatory states increases hepcidin synthesis. Low-calorie and high-protein dietetic interventions have shown an important effect on iron homeostatic regulation. However, these findings and the dietetic source of iron have not been widely explored in people with obesity and iron deficiency. Our study aimed to assess the effect of a low-calorie and high-protein diet with and without red meat on serum hepcidin and iron concentrations in people living with obesity and iron deficiency.

Methods: A randomized clinical trial was performed. Adults with obesity (BMI ≥ 30 Kg/m²), with iron deficiency, and without comorbidities, consumption of medication or iron supplements were included. Participants were randomized to two study groups: a low-calorie and high-protein diet with red meat (RM) and a low-calorie and high-protein diet without red meat (WR), during a two-month follow-up. Baseline and final complete blood count, iron profile test, serum hepcidin levels, biochemical parameters, anthropometry, and body composition analysis were evaluated. A paired-sample t-test or Wilcoxon signed-rank test was used to establish differences between groups. ANOVA was computed to assess differences according to group intervention after follow-up.

Results: Fifty-two participants were included, and forty-seven completed the study (96.2% women). After follow-up, significant weight and fat mass losses were observed (p < 0.001) in both groups, without differences among them. After adjusting for sex, age, and weight loss, we found increased serum iron concentrations, especially among the RM group (p = 0.08). Hepcidin levels rose in both groups after the two-month follow-up period, with a greater increase in the RM group (p = 0.05).

Conclusion: The consumption of a low-calorie and high-protein diet in people living with obesity and iron deficiency seems to have a beneficial effect on weight and fat mass loss, as well as on iron homeostasis, increasing serum levels of iron with a favorable trend among those who consume red meat, but with significant effects on increasing hepcidin levels. More research is needed to determine the mechanism for these findings.

Endashaw Omer, MD, MPH, PNS, ACGF, AGAF¹; Garvit Chhabra, MD²; Abigail Stocker, MD²; Sheel Patel, MD²; Prateek Mathur, MD²; Niang Le, MD²; Carmelita Moppins, APRN²; Lindsay McElmurray, PA-C²; Thomas Abell, MD²; Ethan Steele, DO²; Michael Daniels, MS²

¹University of Louisville, Goshen, KY; ²University of Louisville, Louisville, KY

Financial Support: None Reported.

Background: Intravenous Immunoglobulin (IVIG) therapy is a promising treatment option for patients with drug and device refractory diffuse gastrointestinal (GI) motility disorders, which can severely impact quality of life and lead to nutritional deficiencies. While IVIG has been previously shown to help patients with upper and lower gastrointestinal symptoms, limited research has been done on the potential effects that IVIG has on nutritional status. We report on an ongoing clinical series of patients evaluating the effect of IVIG on nutritional status. We hypothesize that IVIG may improve nutritional parameters such as Body Mass Index (BMI) and Subjective Global Assessment (SGA).

Methods: As part of an ongoing clinical series (NCT04206628), we analyzed data from 142 patients (23 males, 119 females; mean age 49.1 years, Table 1) with GI motility disorders (primarily gastroparesis and functional dyspepsia) treated with IVIG at our institution. Patients were assessed for improvement in nutritional status by measuring BMI and SGA. Data was collected at baseline before starting IVIG and after at least two treatment cycles (one cycle consisting of 12 weeks of weekly IVIG infusions). Data points between groups were compared using t-tests, chi-square tests, or Mann-Whitney U tests, as appropriate. All statistical analyses were performed using R software (version 4.41) and significance was set at p < 0.05.

Results: SGA classification at baseline revealed that 82.4% of patients were in Category B, 16.2% in Category C, and only 1.4% in Category A. Analysis of SGA (Table 1) showed a substantial proportion of patients experienced improvement in their nutritional status. Initially, only 1.4% of the cohort was classified as Category A at baseline. After one treatment cycle, this proportion increased to 6.0%, and by the third cycle, 32.5% of patients were in Category A. Notably, by the 12-month follow-up, 33.1% of patients were in Category A, with most of this improvement coming from those who had moved from category B to A. Figure 1 shows the improvement in SGA with IVIG treatment. Regarding BMI, at baseline, the average BMI for patients was 29.8 ( ± 8.4) which improved to 31.3 ( ± 9.0) (Table 1). As depicted in Figure 2, BMI showed a consistent increase from baseline with each cycle of treatment.

Conclusion: While previous studies have focused on the utility of IVIG therapy for symptom improvement in gastrointestinal dysmotility, these results emphasize that IVIG is a promising treatment modality that also helps improve BMI and overall nutritional status in these patients. Our analysis showed only a few percent of these patients had good nutritional status at baseline (1.4% of patients with SGA Category A) which improved to 33.1% after IVIG. This highlights the effectiveness of treatment regarding nutritional status. An increase in mean BMI was also demonstrated with long-term treatment. Further randomized controlled trials are warranted to better study the role of IVIG therapy in improvement of nutritional status of patients with GI dysmotility, which has always been a clinical challenge.

Table 1. Cohort Characteristics Detailing Trends in SGA and BMI With IVIG Treatments.

Figure 1. Overall Increase in Patients With SGA Category A Compared to Baseline With IVIG Treatment.

Figure 2. Gradual Increase in BMI With IVIG Treatment.

Best of ASPEN-GI, Obesity, Metabolic, and Other Nutrition Related Concepts

International Abstract of Distinction

Giovana Martucelli¹; Danielle Fonseca¹; Ana Prudêncio¹; Dan Linetzky Waitzberg, PhD¹; Raquel Torrinhas, PhD¹

¹Faculty of Medicine of the University of São Paulo, São Paulo, Brazil

Financial Support: CAPES.

Background: Roux-en-Y gastric bypass (RYGB) is a bariatric technique that combines gastric restriction and intestinal malabsorption procedures. It is widely used for managing obesity and type 2 diabetes mellitus (T2D), especially when primary approaches fail. In addition to its effects on weight loss, RYGB may influence gut microbiota (GM), dietary intake (DI), and bile acids (BA), factors that could directly impact the response to T2D remission following surgery. This study aimed to correlate preoperative profiles of GM, DI, and fecal BA in women who were either responders or non-responders to T2DM remission after RYGB.

Methods: This study was a subproject of the thematic study titled The Surgically Induced Metabolic Effects on the Human Gastrointestinal Tract (SURMetaGIT), registered on the Plataforma Brazil and ClinicalTrials.gov. Twenty women from the SURMetaGIT cohort, aged 18 to 60 years, with diagnoses of obesity and T2DM, were included and underwent RYGB. Fecal samples were collected prior to RYGB and used to determine GM by sequencing (Illumina V4 16S rRNA) and assess fecal BA concentrations by mass spectrometry. DI was calculated from a 7-day food diary. One year after the surgery, patients were classified as responders (R) or non-responders (NR) to T2DM remission according to the American Diabetes Association criteria. Correlations were evaluated using Pearson or Spearman tests, with significance set at p ≤ 0.05.

Results: Preoperatively, R and NR women had similar intakes of macronutrients and energy, except for cholesterol intake, which was higher in R. Differences in intestinal bacterial composition preoperatively were characterized by a higher relative abundance of Desulfovibrio piger, Ruminococcus lactaris, Bacteroides nordii and Parabacteroides goldsteinii, and a lower relative abundance of Bacteroides uniformis, Bacteroides salyersiae, and Faecalibacterium prausnitzii in R compared to NR. The bacterial species Ruminococcus lactaris correlated with preoperative cholesterol concentration (r = 0.83; p = 0.041), while Desulfovibrio piger correlated with baseline levels of primary fecal bile acids GCA (r = -0.82; p = 0.023) and TCA (r = -0.77; p = 0.040), only in women who were responders to T2DM remission.

Conclusion: The preoperative profiles of gut microbiota, bile acids, and dietary intake observed in RYGB patients differed according to the glycemic outcome. Specifically, the preoperative correlations between these variables and T2DM remission success suggest that these factors may influence glycemic homeostasis.

Table 1. Dietary Intake of Obese Women Before Roux-En-Y Gastric Bypass, According to the Type of Response to Complete Remission of Type 2 Diabetes Achieved.

Data are expressed as mean ± standard deviation and compared using the Mann-Whitney U test. Comparison in patients with complete remission (R) and without complete remission (NR) of type 2 diabetes: p = NR pre-operative vs. R pre-operative. SFA: saturated fatty acids; MUFA: monounsaturated fatty acids; PUFA: polyunsaturated fatty acids.

Table 2. Fecal bile acid profiles before Roux-en-Y gastric bypass in women who are responders and non-responders to type 2 diabetes remission.

Data are expressed as median [minimum-maximum]. Comparisons in patients with complete remission (R) and without complete remission (NR) of type 2 diabetes: p = NR pre-operative vs. R pre-operative. BA: bile acids; CA: cholic acid; GCA: glycocholic acid; TCA: taurocholic acid; CDCA: chenodeoxycholic acid; GCDCA: glycochenodeoxycholic acid; TCDCA: taurochenodeoxycholic acid; DCA: deoxycholic acid; GDCA: glycodeoxycholic acid; TDCA: taurodeoxycholic acid; LCA: lithocholic acid; GLCA: glycolithocholic acid; TLCA: taurolithocholic acid; UDCA: ursodeoxycholic acid; GUDCA: glycoursodeoxycholic acid; TUDCA: tauroursodeoxycholic acid.

Relative abundance (%) of changes in bacterial sequences before RYGB in women who are non-responders (RSP) or responders (RSC) to T2DM. Results were obtained through DESeq2 analysis, with statistical significance set at p < 0.05.

Figure 1. Intestinal bacterial sequences that showed significant differences preoperatively in Roux-en-Y gastric bypass between women who were responders and non-responders to postoperative remission of type 2 diabetes.

International Abstract of Distinction

Lorena Rodrigues, Msc¹; Andressa Lima, RD¹; Karoline Silva, RD¹; Amanda Santos, RD¹; Luciana Souza, PhD¹; Gabriel Fernandes, PhD²; Joao Mota, PhD¹

¹Federal University of Goias, Goiania, Goias; ²Federal University of Minas Gerais, Belo Horizonte, Minas Gerais

Financial Support: None Reported.

Background: Constipation is the most common digestive complaint among the general population, significantly impacting quality of life. Due to their influence on intestinal motility through modulation of the intestinal microbiota and fermentation processes, probiotics have emerged as a potential treatment for constipation.

Methods: A double-blind, randomized, placebo-controlled clinical trial with a four-week intervention period was carried out over four weeks. Women aged 20 to 59 years, diagnosed with functional constipation, were randomly assigned to either the probiotic group (n = 19), receiving four sachets containing 1x10⁹ CFU Lactobacillus acidophilus (LA-14), 1x10⁹ CFU Lactobacillus casei (LC-11), 1x10⁹ CFU Lactococcus lactis (LL-23), 1x10⁹ CFU Bifidobacterium bifidum (BB-06), and 1x10⁹ CFUBifidobacterium lactis (BL-4), or the placebo group (n = 22) containing 200 mg of maltodextrin per sachet. Dietary intake, physical activity, stool samples, Bristol stool scale, and the Rome IV questionnaire were collected before and after the intervention.

Results: There were no significant differences between the groups in terms of dietary intake, hydration status, physical activity, or anthropometry before and after the intervention. After four weeks of probiotic supplementation, 63.16% of individuals in the intervention group experienced a reversal of constipation, compared to 36.36% in the placebo group (p < 0.05). Severe constipation (Bristol scale type 1) was extinguished in the probiotic group (p < 0.01), whereas it persisted in 27.3% of the placebogroup (p = 0.02). The placebo group exhibited significantly lower microbial diversity (Chao index, p = 0.03) compared to the probiotic group at the end of intervention. No differences were observed in microbial abundances between groups. Participants that improved constipation symptoms showed higher prevalence of Catenibacterium and Enetrorhabdus.

Conclusion: Probiotic supplementation reduced the prevalence of constipation and maintained bacterial diversity. However, further investigation with a larger sample size is warranted to validate these findings.

Table 1. Effect of Probiotic Intervention on Constipation According to the Rome IV Index and Bristol Stool Scale.

Values expressed as absolute numbers (n) and relative values (%) or median (interquartile range)†.

*Difference between baseline and end point. p-value obtained by Wilcoxon test or Fisher's Exact test.

p-value obtained by the Z-test for two proportions, independent samples.

Figure 1. Percentage Variation (post – pre-intervention, Δ) of Individuals Classified With Constipation According to the Rome IV Criteria.

A: placebo group and B: probiotic group. *p = 0.03.

Figure 2. Assessment of Gut Microbiota Diversity.

International Abstract of Distinction

Qian Ren, PhD¹; Jinrong Liang²; Peizhan Chen, Clinical Research Center, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine³

¹Department of Clinical Nutrition, Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai; ²Department of Oncology, Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai; ³Clinical Research Center, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai

Financial Support: This study was supported by the Youth Cultivation Program of Shanghai Sixth People's Hospital (Grant No. ynqn202223), the Key Laboratory of Trace Element Nutrition, National Health Commission of the Peoples’ Republic of China (Grant No. wlkfz202308), and the Danone Institute China Diet Nutrition Research and Communication (Grant No. DIC2023-06).

Background: To investigate whether a high-fat diet (HFD) can induce sarcopenic obesity (SO) phenotype and the underlying mechanisms.

Methods: Five-week-old male C57BL/6J mice (n = 12 per group) were randomly divided into the Control group (Con, AIN-93G diet) and the HFD group (HFD, isocaloric, with 60% percentage of energy from fat), which were investigated for 12 weeks.

Results: Following the 12-week intervention, body weight of the HFD group significantly increased compared to the Con group (Figures 1A-B). Concurrently, there was a significant increase in body fat percentage and a significant decrease in lean tissue percentage, as assessed by MRI (Figures 1C-E). H&E staining of the gastrocnemius muscle in the HFD group mice indicated adipose infiltration (Figure 1F). RNA-seq analysis identified a total of 349 differentially expressed genes in the mouse skeletal muscle tissue, of which 203 were upregulated and 146 were downregulated (Figures 2A-D). Validation of some differentially expressed genes through qPCR and Western blotting revealed that the expression of the circadian gene Per2 at the mRNA level (Figure 2E) and protein level (Figure 2F) in the gastrocnemius muscle tissue of the SO model mice (HFD group) was significantly upregulated compared to the normal mice (Con group).

Conclusion: A high-fat diet may increase the risk of SO by inducing the loss of skeletal muscle mass and the increase of adipose tissue through the upregulation of Per2 expression.

Figure 1. Following the 12-week intervention, body weight (Figures 1A-B), body composition assessed by MRI (Figures 1C-E) and H&E staining of the gastrocnemius muscle (Figure 1F) of the HFD group and the Con group mice.

Figure 2. RNA-seq analysis in the mouse skeletal muscle tissue (Figures 2A-D), and validation of Per2 mRNA and protein expression through qPCR (Figure 2E) and Western blotting (Figure 2F) in the gastrocnemius muscle tissue of the mice in HFD group and Con group.

Pediatric, Neonatal, Pregnancy, and Lactation

Best of ASPEN-Pediatric, Neonatal, Pregnancy, and Lactation

Mirielle Pauline, PhD, BSc¹; Rohan Persad²; Pamela Wizzard, BSc, RAHT²; Evan Labonne²; Mahabub Alam, MSc, DVM²; Patrick Nation, DVM²; Justine Turner, PhD, MD²; Paul Wales, MD³

¹University of Alberta, St. Albert, AB; ²University of Alberta, Edmonton, AB; ³Cincinnati Children's Hospital Medical Center, Cincinnati, OH

Financial Support: None Reported.

Background: Surgical causes of short bowel syndrome (SBS) in neonates frequently results in interrupted bowel continuity with an initial diverting stoma. Hence, the distal intestinal remnant is usually not used until continuity is restored at a subsequent operation and does not have exposure to nutrition, proximal intestinal secretions or proximal trophic factors. Nutrition and trophic factors are key drivers of intestinal adaptation that is necessary for patients to achieve autonomy from parenteral nutrition (PN). A novel clinical strategy has emerged using distal bowel refeeding (DBR), taking proximal effluent from the stoma and refeeding it via an enteric fistula into the distal remnant intestine. Case studies suggest DBR reduces PN volumes, increases enteral nutrition (EN) tolerance, improves the quality of distal intestinal tissue and may reduce complications of PN. The aim of this pilot study was to develop an experimental model of DBR in neonatal piglets with SBS.

Methods: Neonatal piglets aged 2-3 days underwent this novel experimental model of SBS with a 75% distal small bowel resection with jejunostomy, leaving 25% of the proximal bowel, 10 cm of terminal ileum and an intact colon. A gastrostomy tube was inserted into the stomach for decompression and enteric tubes were placed into the remnant proximal jejunum (functional stoma) and into the terminal ileum (enteric fistula) for DBR. A jugular catheter was placed for 100% PN delivery and for fluid supplementation as required. Piglets were randomized to DBR every 4 hours starting day 2 (DBR, n = 6) or control with 3mL saline delivered into the distal bowel at equivalent timepoints (CON, n = 4). Stoma output was recorded every 4 hours as part of careful fluid balance monitoring. After 7 days, blood was collected for liver chemistry, electrolytes and measurement of short chain fatty acids (SCFAs). At laparotomy, small intestinal length and weight, pre-stoma jejunum diameter and post-stoma ileum diameter were all measured. Jejunal tissue was collected for histology and gene expression via real-time quantitative polymerase chain reaction (qPCR) (data pending). Data is analyzed via Mann-Whitney U Test and presented as median (interquartile range).

Results: At initial surgery there was no difference between piglet groups age (p = 1.0), weight (p = 0.07), pre-resection intestinal length (p = 0.48) or post-resection length (p = 0.35) (Table 1.). Over the course of the trial average daily stoma fluid output was reduced for DBR compared to CON (p = 0.02) (Figure 1). Day 7, DBR piglets gained more weight from baseline (p = 0.038) and weighed more (p = 0.01). The distal bowel diameter was increased for DBR over CON (p = 0.01). There was no difference in liver chemistry, electrolytes, plasma SCFAs or jejunal scraping weight.

Conclusion: This pilot study introduces a new neonatal piglet model of DBR in SBS. Piglets who received DBR demonstrated improved weight gain, decreased stoma fluid losses and increased distal bowel diameter, which would permit re-anastomosis with minimal size discrepancy. In this small pilot study of short duration, without weaning of TPN, we found no benefit on liver chemistry or a gross measure of adaptation (jejunal weight), although jejunal histology is pending. The mechanisms for reduced proximal stoma fluid losses with DBR warrant further investigation and we are currently exploring the impact of DBR on distal gut trophic factor expression.

Table 1. Adaptation Metrics.

Figure 1. Average jejunum output per day. Analyzed using Mann-Whitney U Test.

Paul Wales, MD¹; Beth Carter, MD²; Valeria Cohran, MD³; Susan Hill, MD⁴; Samuel Kocoshis, MD⁵; Brian Terreri, PharmD⁶; Sharif Uddin, MS⁶; Robert Venick, MD⁷; Danielle Wendel, MD⁸

¹Cincinnati Children's Hospital Medical Center, Cincinnati, OH; ²Children's Hospital Los Angeles, Los Angeles, CA; ³Ann & Robert H. Lurie Children's Hospital of Chicago, Chicago, IL; ⁴Great Ormond Street Hospital for Children NHS Foundation Trust, London, England; ⁵Cincinnati Children's Hospital Medical Center, Cincinnati, OH; ⁶Takeda Pharmaceuticals USA, Inc., Lexington, MA; ⁷UCLA Mattel Children's Hospital, Los Angeles, CA; ⁸Seattle Children's Hospital, Seattle, WA

Financial Support: Takeda Pharmaceuticals U.S.A., Inc.

Background: Short bowel syndrome-associated intestinal failure (SBS-IF) is a rare malabsorption disorder in which the length of functional bowel is reduced often as a result of intestinal resection. In children, diarrhea caused by SBS-IF can lead to malnutrition and impair quality of life, increasing the need for parenteral nutrition and/or intravenous fluids (PN/IV) to maintain health and growth. Teduglutide is a glucagon-like peptide-2 analog that can reduce dependence on PN/IV. This post hoc analysis evaluated the impact of teduglutide on stool characteristics and PN/IV use across 96 weeks in pediatric patients with SBS-IF.

Methods: Data were pooled from two open-label, multicenter, prospective, phase 3, long-term extension studies of teduglutide in pediatric patients with SBS-IF (NCT02949362, NCT02954458) who completed the core trials. Patients received teduglutide (0.05 mg/kg/day) in 24-week treatment cycles followed by a 4-week, no-treatment follow-up period for up to 6 cycles. Data are presented here up to cycle 4, week 24 (96 weeks of teduglutide treatment). The objective of this analysis was to evaluate stool characteristics including Bristol Stool Form Score (BSFS; a 7-point scale used to evaluate stool consistency from 1, representing ‘hard lumps’, to 7, representing ‘watery’) and number of stools per day in addition to PN/IV-related outcomes from the start of the studies (cycle 1, week 1) to 96 weeks of treatment (cycle 4, week 24). Descriptive statistics are presented here.

Results: In total, 69 patients were included at the start of the studies; mean (standard deviation [SD]) patient age was 6.8 (3.75) years, and the majority of patients (66.7%) were male. The most common reason for resection or diagnosis was gastroschisis (36.2%) followed by midgut volvulus (29.0%) and necrotizing enterocolitis (17.4%). Colon-in-continuity was present in 93.8% of patients with remaining colon. In total, 20.3% of patients had a stoma (jejunostomy: 64.3%; ileostomy: 14.3%; colostomy: 14.3%; other: 7.1%). Mean BSFS decreased between the start of the studies and 12 weeks of treatment; fluctuations in mean BSFS were observed between 12 and 96 weeks (Figure 1). Mean (SD) stool frequency was 3.8 (2.94; n = 51) stools per day at the start of the studies and 3.3 (1.93; n = 24) stools per day at 96 weeks (Table 1). PN/IV-related outcomes improved from the start of the studies to week 96 of treatment. Mean (SD) PN/IV volume decreased from 48.5 (33.04; n = 68) mL/kg/day at the start of the studies to 31.8 (29.19; n = 33) mL/kg/day at week 96 of treatment (Table 2). Overall, the mean number of days patients required PN/IV every week decreased over time from the start of the studies to week 96 of treatment (Figure 2). Mean (SD) PN/IV duration decreased from 9.6 (4.40; n = 68) hours per day at the start of the studies to 7.3 (5.88; n = 33) hours per day at week 96 of treatment. Of those with a stoma, mean ostomy output increased over time; mean (SD) output was 45.0 (27.29; n = 9) mL/kg/day at the start of the studies and 49.9 (30.60; n = 6) mL/kg/day at week 96 of treatment.

Conclusion: This post hoc analysis of long-term teduglutide treatment suggests a trend towards improved stool consistency and frequency in pediatric patients with SBS-IF receiving teduglutide for 96 weeks. This was accompanied by a reduction in PN/IV requirements in terms of volume and frequency. A limitation of these observations is the small and reducing sample sizes over time. These findings could inform clinicians regarding the potential benefits of teduglutide for managing pediatric patients with SBS-IF. Research is warranted to evaluate whether there is an association between improvements in stool characteristics and quality of life after teduglutide treatment.

Table 1. Mean (SD) and Median (range) Number of Stools per Day Over Time.

Table 2. Mean (SD) and Median (range) PN/IV Volume (mL/kg) per Day Over Time.

Figure 1. Mean (SE) Bristol Stool Form Score per Week Over Time.

Figure 2. Mean (SE) Number of Days Receiving PN/IV per Week Over Time.

Anam Bashir, MBBS¹; Mary Bridget Kastl, MSN, RN, CRNP, FNP-BC¹; Laura Padula, MS, RD, LDN¹; Elizabeth Reid, MS, RDN, LDN¹; Rachel Kofsky, RD¹; Maria R. Mascarenhas, MBBS¹

¹Children's Hospital of Philadelphia, Philadelphia, PA

Financial Support: None Reported.

Background: Cystic fibrosis (CF) is a genetic disorder that primarily affects the lungs and digestive system, leading to a range of nutritional challenges and complications. Historically, malnutrition has been a significant concern among people with CF (PwCF). However, recent advancements in treatment, particularly the introduction of cystic fibrosis transmembrane conductance regulator (CFTR) modulators, have improved pulmonary outcomes and led to a noticeable increase in weight gain and BMI z scores in this population. This study aims to evaluate the prevalence of overweight and obesity among patients with cystic fibrosis and assess the comorbidities in this population.

Methods: A retrospective chart review assessed PwCF (2-23 years of age) at the Children's Hospital of Philadelphia. Data collected included demographics, CF genotype, anthropometric measurements (height, weight, and body mass index), pancreatic function (assessed through fecal elastase levels), medication history, and any comorbid diagnosis. Patients were classified based on their body mass index (BMI): underweight as BMI < 5th percentile, normal weight as BMI 5th - 85th percentile, overweight as BMI 85th - 95th percentile, and obese as BMI > 95th percentile.

Results: A total of 243 patients with cystic fibrosis, with a mean age of 10.4 years (53% male), are currently followed at our center. Within this cohort, 4 patients (1.6%) were classified as malnourished, 192 patients (79%) with normal weight, and 47 patients (19%) as overweight/obese (OW): 26 patients (10.6%) were overweight and 21 patients (8.6%) were obese. The mean age of OW group was 10.5 years (61.7% male). Of the OW patients, 21 patients (44%) carried two severe mutations, 26 (56.5%) were pancreatic insufficient, and 40 (85%) were receiving CFTR modulator treatment. Of the OW patients on CFTR modulators, 52.5% (n = 21) were classified as OW after being on CFTR modulators for a mean duration of 17 months, while 47.5% (n = 19) were already OW prior to initiating CFTR modulator therapy. In the OW population, 4 patients (8.5%) were diagnosed with cystic fibrosis-related diabetes (CFRD), 2 (4.2%) with impaired glucose tolerance, 4 (8.5%) with obstructive sleep apnea, and 1 (2%) with hyperlipidemia.

Conclusion: The prevalence of overweight/obesity in people with CF is high (18.3%), with a male predominance. Overweight/obesity is prevalent even in those with severe phenotypes and pancreatic insufficiency. Over half of OW patients on CFTR modulators develop overweight/obesity after starting CFTR modulator therapy. Despite the high prevalence of OW status in people with CF, the prevalence of comorbidities is low.

Abstract of Distinction

Elias Wojahn, BS¹; Liyun Zhang, MS²; Amy Pan, PhD²; Theresa Mikhailov, MD, PhD²

¹Medical College of Wisconsin, Wauwatosa, WI; ²Medical College of Wisconsin, Milwaukee, WI

Financial Support: None Reported.

Background: Adequate nutrition is crucial for the well-being and healthy development of children, including when attempting to recover from illness. Sepsis and septic shock are life-threatening conditions that affect all systems of the body. The widespread release of cytokines and hormones increases metabolic demand, precipitating additional risk from malnutrition in these patients. Previous work has shown the benefit of early nutrition in critically ill patients. The aim of this study is to elucidate the nature of the association between timely screening for malnutrition in pediatric sepsis and septic shock patients and clinical outcomes. We hypothesize that patients with malnutrition had higher mortality rates and longer lengths of stays compared to those who did not.

Methods: We retrospectively obtained patient information regarding diagnoses, screenings, and outcomes from October 2019 to December 2023 for 18 sites participating in the optional Nutrition Module from Virtual Pediatric Systems, LLC (VPS). We categorized each patient as malnourished or non-malnourished according to the results of the nutrition screen at admission to the PICU. We examined clinical outcomes, specifically mortality and PICU and hospital length of stay (LOS). Categorical variables were compared via Chi-square test and continuous variables via Mann-Whitney-Wilcoxon test. We then performed multivariable analysis via logistic regression and/or a general linear model to control for severity of illness, age group, sex assigned at birth, race/ethnicity, trauma and patient type.

Results: We found 1610 pediatric patients with septic shock who were screened for malnutrition. 480 (29.8%) were at risk, 1130 (70.2%) were not. Malnutrition was not associated with sex or age-group but there was a difference by race/ethnicity (p < .0001). Mortality did not differ between those who were malnourished and those who were not (6.25% vs 5.22%, p = .41). This relationship persisted when controlling for age group, gender, race/ethnicity, trauma, patients’ type and PRISM3 (p = .39). However, malnutrition was associated with a significantly longer PICU LOS [3.87 (1.76-9.66) vs 2.57 (1.20-6.41), median (IQR) in days p < .0001] and hospital LOS [11.78 (5.58-25.15) vs 7.16 (3.89-15.64), median (IQR) in days, p < .0001)]. This remained true even after controlling for the potential confounders (p < .0001).

Conclusion: Mortality did not differ between pediatric patients with sepsis or septic shock who screened as malnourished and those who did not. Pediatric patients with sepsis or septic shock who screened as malnourished had longer LOS in the PICU and the hospital than pediatric sepsis or septic shock patients who did not. Thus, early screening for malnutrition may provide an opportunity to impact clinical care and should be initiated to minimize the emotional and financial costs associated with prolonged length of stays.

Abstract of Distinction

Katie Huff, MD, MS¹; Zhihong Yang, PhD¹; Suthat Liangpunsakul, MD, MPH¹

¹Indiana University School of Medicine, Indianapolis, IN

Financial Support: None Reported.

Background: Patients with intestinal failure rely on parenteral nutrition (PN) for survival. However, PN is associated with complications, including intestinal failure-associated liver disease (IFALD). Direct bilirubin levels are traditionally used to diagnose and monitor IFALD, but they can normalize even with ongoing liver disease. This study aimed to identify differences in gene and metabolite expression in neonates with and without IFALD.

Methods: Neonates requiring PN for more than 2 weeks, who had received less than 72 hours of PN, and had no baseline liver disease were eligible. IFALD was defined by a direct bilirubin level >2 mg/dL during PN. Serial blood samples were collected at PN initiation (baseline) and at the development of IFALD or the end of PN (final). RNA sequencing identified differentially expressed messenger RNA (mRNA) transcripts, with a false discovery rate < 0.05 and absolute fold change ≥ 2 considered significant. Untargeted primary metabolomics using gas chromatography-mass spectrometry (MS) and lipidomics using quadrupole time-of-flight MS/MS were performed. Metabolite peak intensities were compared using MetaboAnalyst software, with a significance threshold of p-value < 0.1.

Results: Fourteen subjects were included (7 with IFALD, 7 without or non-IFALD), and demographic information is provided in Table 1. When comparing all four groups (baseline IFALD, baseline non-IFALD, final IFALD, and final non-IFALD), there were no significant differences in metabolites from primary metabolomics or lipidomics from Partial least squares-discriminant analysis (PLS-DA). However, there were significant differences in overall metabolite profiles when comparing baseline to final samples (Figure 1). Primary metabolomic analysis revealed multiple metabolites with significantly altered concentrations (p < 0.1), as outlined in Table 2. In lipidomics, six triglyceride types with increased concentrations (p < 0.1) were identified at final sampling in the IFALD group. At final sampling, five genes were differentially expressed between groups, with four downregulated and one upregulated in the IFALD group (p < 0.001, Figure 2). Notably, four of these genes are associated with the spliceosome, a pathway previously linked to hepatic steatosis in non-alcoholic fatty liver disease.

Conclusion: Differences in gene and metabolite expression were observed in neonates receiving PN who developed IFALD. Many of these altered metabolites and genes have been previously linked to liver disease. Further research is needed to explore the role of additional patient factors, including specific PN components, and their impact on outcomes. Additionally, monitoring these metabolites over time could provide insights into the progression of IFALD.

Table 1. Demographic and Patient Information Compared Between Outcome Groups.

(All data presented as mean ± stdev unless noted, Mann-Whitney U test used for analysis).

Table 2. Metabolites from Primary Metabolomic Analysis That Were Significantly Different Between Groups.

All metabolites listed had a p-value < 0.1 when concentrations compared between IFALD and non-IFALD groups.

A. PLS-DA score plot of metabolomic differences between the four groups, each dot represents a sample and each ellipses 95% of patient samples. B. Heatmap of metabolite differences between groups with each column representing a patient sample and each row a metabolite. C. PLS-DA score plot of lipidomic differences between groups. D. Heatmap of lipidomic differences between groups.

Figure 1. Comparison of Groups Prior to and After Parenteral Nutrition Exposure With Primary Metabolomics and Lipidomics.

Comparing IFALD (black) and non-IFALD (gray) groups all mRNA levels were significantly different between groups with p-value < 0.001.

Figure 2. Differentially Expressed Genes at Final Sampling.

Abstract of Distinction

Caitlin Bowers, BA¹; Stephanie Merlino Barr, PhD, RDN, LD²

¹Case Western Reserve University School of Medicine, Cleveland, OH; ²MetroHealth Medical Center, Cleveland, OH

Financial Support: None Reported.

Background: Fat free mass accretion among premature, very low birthweight (VLBW) infants is associated with organ development and improved neurodevelopmental outcomes. Nutrition interventions may play a role in fat free mass accretion. Adjusted fat free mass z-scores measure body composition development while correcting for approximated lean body mass at birth. This study investigated the relationship of first week nutrient intake and total human milk consumption with body composition development.

Methods: This was a single-center, retrospective cohort study performed among premature, VLBW (birthweight < 1500 g) patients admitted to a level III neonatal intensive care unit (NICU) from 2018–2024 who received body composition analysis via air displacement plethysmography. Fat free mass z-scores were calculated using the Norris 2019 body composition growth charts. Adjusted fat free mass z-score was calculated as the difference between fat free mass z-score and birthweight z-score. Welch's t test, Wilcoxon rank sum test, and chi square tests were performed to compare infants who consumed majority mother's own milk (MOM) with other infants. Multiple linear regressions were performed to assess the effect of average first week caloric intake (kcal/kg/day), average first week protein intake (g/kg/day), average first week non-protein energy (NPE) to protein ratio, and total human milk consumption (mL) over the NICU course on adjusted fat free mass z-score. All models controlled for birth gestational age (weeks), sex (male vs. not male), and maternal race (Black/African American vs. not). Models evaluating human milk intake also controlled for length of NICU stay (days). Identified variables (see Table 2) were log transformed to meet assumptions of normality. Missing values were imputed using polytomous regression. Analyses were performed using R version 4.2.

Results: 175 infants were included in the study. Descriptive characteristics and comparative tests are reported in Table 1. Infants who consumed majority MOM were born less prematurely, more often female, and trended towards a shorter NICU stay. Adjusted fat free mass z-scores displayed a narrower distribution and a less negative mean compared to fat free mass z-scores (Figure 1). While an inverse relationship was observed between fat free mass z-score and birth gestational age, no relationship between adjusted fat free mass z-score, birth gestational age, and total MOM intake was visually observed (Figure 2). Multiple linear regression models are reported in Table 2. Average caloric intake over the first week of life was associated with a decrease in adjusted fat free mass accretion (β = -0.01, 95% CI = -0.02, 0.00, p = 0.034). Average first week protein intake and NPE to protein ratio were not significant predictors of adjusted fat free mass z-score. Similarly, total human milk, MOM, and donor milk intake were not significant predictors of adjusted fat free mass z-score.

Conclusion: Fat free mass accretion is associated with improvements in neurodevelopment in preterm infants. While past research has proposed human milk intake improves lean body mass accretion, this relationship was not observed in the present study. Continued research is warranted to better understand nutrition's role in quality of growth of preterm infants. These findings support the use of adjusted fat free mass z-score to evaluate nutritional interventions. Adjusted fat free mass z-score corrects for approximated body composition at birth to allow for evaluation of growth over the postnatal period. Further research should investigate the connection between fat free mass accretion in the postnatal period and clinical outcomes of preterm infants.

Table 1. Clinical and Demographic Characteristics of Patients by Mother's Own Milk (MOM) Intake.

1. Non-normally distributed variable, median and interquartile range reported, significance determined by Wilcoxon rank sum test. 2. Normally distributed variable, mean and standard deviation reported, significance determined by Welch's t test. 3. Significance by chi square test. 4. Race missing for 9 patients. 5. Ethnicity missing for 2 patients.

Table 2. Multiple Linear Regression Models Predicting Adjusted Fat Free Mass Z-Score.

*p < 0.05, **p < 0.01, ***p < 0.001.

Figure 1. Density Plots of Fat Free Mass Z-Score and Adjusted Fat Free Mass Z-Score.

Figure 2. Fat Free Mass Z-Score and Adjusted Fat Free Mass Z-Score Versus Gestational Age at Birth With Total Mother's Own Milk Intake.