Abstract
Objectives:
To determine the impact of free water administration on clinical outcomes in medically complex patients, based on the hypothesis that patients receiving more thin liquids have worse outcomes related to increased risk of gastroesophageal reflux.
Methods:
This was a retrospective chart review of children initiated on commercial blenderized tube feedings from 2010 to 2019. The percentage of gastrostomy intake that was a thin liquid was determined, with thin liquids including free water or thin formula based on the International Dysphagia Diet Stardardization Initiative (IDDSI) framework. Patients were categorized into those receiving low volumes of thin liquids (“percent thin” < 20%) and those receiving higher volumes (“percent thin” > 20%). Emergency room visits, hospital admissions, and chest x-rays during the year after starting the blend were compared.
Results:
Forty-five patients age 12 months to 18.7 years (median 3.7 years) were included. Twenty patients (44%) were receiving a blend with an IDDSI-categorized thin consistency and 25 patients (56%) were receiving a thicker blend. In addition to the blend, patients received a median of 320mL per day of water (range 0–1000mL). Patients receiving < 20% thin liquids were less likely to undergo chest x-rays during follow-up than patients receiving larger amounts of thin liquids (10% in the minimal thin group vs. 48% in the greater thin group, P = 0.03). In a multivariable logistic regression, this relationship remained significant after controlling for underlying pulmonary disease, aspiration, method of feed administration (bolus or continuous feeds), fundoplication status, and oral intake status.
Conclusions:
Our study demonstrates that patients on blenderized tube feeds receive widely variable amounts of added thin liquids, like water. The addition of water to tube feeds, while typically given for hydration, can modify feed viscosity and clinical outcomes, such as chest x-ray performance.
Keywords: Feeding, blenderized feeds, aspiration, reflux
Introduction:
Blenderized tube feeds have emerged as a common alternative to conventional formula for patients with feeding difficulties.1 In addition to homemade blenderized feeds, there are a growing variety of commercially available blend formulas that have similar properties to homemade blends. Blenderized feeds have many benefits and have been shown to reduce symptoms like retching and vomiting, decrease respiratory disease, improve quality of life, and reduce costs.2–7 One proposed mechanism for positive gastrointestinal and respiratory outcomes relates to the viscosity of blends, which tends to be higher than typical commercial formulas.8,9 Higher viscosity feeds move more quickly to the antrum of the stomach, which may reduce reflux and reflux-associated gastrointestinal (GI) and respiratory symptoms.10–12
Although the viscosity of blenderized feeds tends to be higher than conventional formulas, feed viscosity varies based on the brand of commercial feeds used and the amount of water added to the blends, which can lead to thinning of an otherwise thick formula.9 No prior studies have examined the impact of free water administration or lower viscosity feeds on clinical outcomes in patients receiving blenderized formulas. Thus, the aims of the current study were to investigate the impact of receiving thin liquids, including free water, on hospital admissions, emergency room visits, and chest x-ray performance in patients receiving commercial blenderized tube feeds. We hypothesized that patients receiving less thin liquids would have lower rates of hospitalization and emergency room visits and improved respiratory outcomes as evidenced by reduced chest x-ray performance.
Methods:
This was a retrospective study of patients started on a commercial blenderized formula at Boston Children’s Hospital from 2010 to 2019. A search was conducted for prescriptions for the following commercial blenderized formulas: Compleat Pediatric and Compleat Pediatric Organic Blends (Nestlé Health Science, Lausanne, Switzerland), Kate Farms (Kate Farms, Santa Barbara, CA), Liquid Hope and Nourish (Nutritional Medicinals, LLC, Westchester, OH), Pediasure Harvest (Abbott Nutrition, Chicago, IL), and Real Food Blends (Real Food Blends, Chesterton, IN). Inclusion criteria included patients who were newly started on a commercial blenderized formula. To minimize feeding plan heterogeneity from patients receiving large amounts of other formulas or solid foods, patients were considered for inclusion only if at least 50% of total intake was the commercial blend. Patients were excluded if they were receiving feeds via gastrojejunostomy tube, if they never started the formula, if there was lack of follow-up, if the chart contained minimal information about the blend or free water regimen, or if the patient was receiving parenteral nutrition.
Institutional review board approval was obtained for this study. Study data were collected and managed using REDCap electronic data capture tools hosted at Boston Children’s Hospital.
Data Collection
Patient records were reviewed for baseline data, including age, sex, weight-for-age z-score, GI diagnoses, comorbidities, surgeries, enteral tube type, videofluoroscopic swallow studies, feeding regimen, and free water regimen. Comorbidities were characterized as pulmonary, cardiac, neurologic, metabolic or genetic, endocrine, immunologic, oncologic, oropharyngeal malformation, or prematurity (defined as gestational age of <37 weeks). Data on the feeding regimen included the type of commercial blenderized formula initiated and the method of administration (bolus, continuous, or combination). Additional purees, if received, also were noted. Blenderized formulas were divided into 5 viscosity categories as follows: Thin (Compleat Pediatric and Kate Farms), Slightly thick (Harvest), Mildly Thick (Nourish), Moderately Thick (Compleat Pediatric Organic Blends and Liquid Hope), Extremely Thick (Real Food Blends).9 The total amount of free water per day also was recorded.
Each patient’s estimated total fluid needs were calculated based on the Holliday-Segar method, and the percent of estimated fluid needs received from all sources was calculated. Patients were characterized based on the percentage of all fluid they received that was a thin liquid (called “percent thin”). The “percent thin” was defined as the patient’s daily volume of water plus the patient’s daily volume of thin blend (Compleat Pediatric or Kate Farms) divided by the patient’s total daily fluid volume. Patients were stratified into those that received minimal thin liquids (“percent thin” < 20%) and those receiving higher amounts of thin liquids (“percent thin” > 20%) based on this being the lowest quartile in the study population.
Outcomes
Outcome measures were recorded one year after the initiation of the commercial blend. All charts were reviewed to ensure that the patient remained on gastrostomy feeds during the year of follow-up without a change in formula and without a change from bolus to continuous feeds. Outcomes included whether the patient had any emergency room visits or any hospitalizations during the follow-up period. To assess respiratory-related outcomes and concerns, we recorded whether any chest x-rays were received during the follow-up year. Chest x-ray findings were categorized as normal or containing atelectasis, consolidation(s), or prominent interstitial markings. To assess hydration status, the chart was reviewed for chemistries or urine specific gravities obtained over the follow-up year while the patient was well.
Statistical Analyses
Continuous variables are expressed as means with standard deviations, and categorical variables are expressed as n (%). Comparisons of continuous variables were completed using t tests. Comparisons of categorical variables were completed using χ2 tests or Fisher exact tests. The percentage of patients with any emergency room visits, hospital admissions, chest x-rays, or abnormal chemistries during the follow-up period were compared between the minimal thin (“percent thin” < 20%) and higher percent thin (“percent thin” > 20%) groups. Given a statistically significant result for chest-rays, a multivariable logistic regression was performed to assess for factors affecting whether patients required any chest x-rays during the follow-up period. All tests were two-sided with P < 0.05 considered statistically significant. Odd ratios are expressed with 95% confidence intervals (CI). Statistical analysis was conducted using Stata (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC).
Results:
We identified 102 patients who were prescribed a commercial blenderized formula for enteral tube support for the first time between 2010 and 2019, and 79 of these patients were taking at least 50% of their intake by volume through these formulas. Thirty-four patients were excluded due to the patient receiving gastrojejunostomy feeds (n = 5), the patient never starting the formula (n = 6), lack of follow-up data after starting the formula (n = 4), lack of detail about the blend or free water regimens (n = 13), or concurrent receipt of parenteral nutrition (n = 6). Forty-five patients were included in the final analysis.
The demographic and clinical characteristics of the study population are shown in Table 1. The median age at initiation of the blend was 3.7 years (range 12 months to 18.7 years). Patients had an average of 3.2 ± 1.5 non-gastrointestinal comorbidities, with the most common being neurologic (76%) and pulmonary (64%). Patients had 2.9 ± 1.2 GI diagnoses at the time of starting the blend, with the most common being constipation (62%), reflux (56%), and aspiration (51%). All patients in the study had gastrostomy tubes. Thirty-six patients (80%) were receiving bolus feeds, 3 (7%) were receiving continuous feeds, and 6 (13%) were receiving a mix of bolus and continuous feeds. All patients were receiving at least 90% of daily volume via enteral tube. Twenty-nine patients (64%) were taking small sips or bites orally in addition to the tube feeds.
Table 1.
Patient demographics and clinical characteristics
| Age in years, median (range) | 3.7 (1.0–18.7) |
| Male | 28 (62%) |
| Weight z-score, mean ± SD | −1.1 ± 1.3 |
| Comorbidities | |
| Neurologic | 34 (76%) |
| Pulmonary | 29 (64%) |
| Metabolic/Genetic | 22 (49%) |
| Cardiac | 13 (29%) |
| Oropharyngeal malformation | 9 (20%) |
| Prematurity | 12 (27%) |
| GI Diagnoses | |
| Constipation | 28 (62%) |
| Reflux | 25 (56%) |
| Aspiration | 23 (51%) |
| Vomiting and/or retching | 20 (44%) |
| Laryngeal penetration | 5 (11%) |
| Diarrhea | 4 (9%) |
| EA/TEF | 3 (7%) |
| Surgeries | |
| Gastrostomy tube | 45 (100%) |
| Fundoplication | 7 (16%) |
| • Intact | 6 (13%) |
| • Unwrapped | 1 (2%) |
| EA/TEF repair | 3 (7%) |
| Formula | |
| Compleat Pediatric* | 19 (42%) |
| Compleat Pediatric Reduced Calorie* | 1 (2%) |
| Nourish | 10 (22%) |
| Liquid Hope | 1 (2%) |
| Compleat Pediatric Organic Blends | 11 (24%) |
| Real Food Blends | 3 (7%) |
EA, esophageal atresia; TEF, tracheoesophageal fistula
denotes a thin liquid formula
Patients were stratified according to the viscosity of their blendarized formula. Twenty patients (44%) received a blend with a thin consistency (Compleat Pediatrics or Compleat Pediatric Reduced Calories) and 25 patients (56%) received a thicker blend (Nourish, Liquid Hope, Compleat Pediatric Organic Blends, or Real Food Blends). In addition to their blend, patients received a median of 320mL per day of additional water (range 0–1000mL). The amount of added water in addition to formula was similar regardless of the viscosity of the commercial blend being used (Figure 1). In addition to formula and water, 17 patients (38%) were receiving puree by tube as a form of hydration. Through blends, water, and purees, patients received an average of 84.3 ± 19.5 percent of estimated total free fluid needs based on the Holliday-Segar method.
Figure 1:
Comparison of the proportion of formula and added water in patients on thin and thick viscosity commercial blends (P = 0.46)
To assess the effect of thin liquids in the form of additional free water or a thin blend on patient outcomes, patients were characterized based on the percentage of total daily fluids that were a thin liquid (“percent thin”). The median percent thin was 47 (range 0–100%). Ten patients received minimal thin liquids, with a percent thin < 20%, and 35 patients received >20% of thin liquids. There were no differences in the underlying comorbidities between the minimal thin and higher thin group, including in incidence of constipation (P > 0.05). The percent of estimated total fluid needs received was similar between groups (77.3 ± 16.2 in the <20% thin group vs. 86.3 ± 19.5 in the >20% thin group, P = 0.20), and they received the same amount of total fluid per kilogram per day (69 cc/kg/day in the <20% thin group vs. 69 cc/kg/day in the >20% thin group, P = 0.92).
Figure 2 shows the percentage of patients with any emergency room visits, hospital admissions, chest x-rays, or abnormal chemistries compared between the minimal thin and higher thin groups. Patients in the minimal thin group were less likely to undergo chest x-rays during the follow-up year when compared to the higher thin group (10% of patients received chest x-rays in the minimal thin group vs. 49% in the higher thin group, P = 0.03). Logistic regression was performed to assess for factors affecting whether patients required a chest x-ray during the follow-up period. When pulmonary comorbidities, aspiration, fundoplication status, feeding method (bolus vs. continuous feeds), and oral feeding status (oral feed vs. no oral feeds) were added as covariates, receiving minimal thin liquids remained independently associated with decreased number of chest x-rays (P = 0.04). Having a pulmonary comorbidity also was associated with increased number of chest x-ray (P = 0.03). Overall, 4/20 (20%) patients had chest x-rays that were normal, 3/20 (15%) had atelectasis, 6/20 (30%) had consolidations, and 7/20 (35%) had prominent interstitial markings. Only one patient in the minimal thin group had a chest x-ray, and this showed prominent interstitial markings.
Figure 2:
Effect of the proportion of thin feeds on patient outcomes
As a measure of hydration status, chemistries and urine specific gravities during a period of wellness were recorded. Twenty-two patients (49%) had chemistries, and 41% of these were abnormal. Abnormalities included low bicarbonate, sodium, or chloride levels. These abnormalities were more frequent in the >20% thin group than the <20% thin group, although this was not statistically significant (44% of patients in the >20% thin group had abnormal chemistries vs. 25% in the < 20% thin group, P = 0.61). Only 2 patients had values for urine specific gravity while the patient was well; these were 1.021 and 1.023.
Discussion
This is the first study to examine the impact of thin liquids including free water administration on clinical outcomes in patients receiving commercial blenderized tube feeds. Water and other thin liquids are routinely added to enteral feeds for hydration, however the decision to add water is rarely considered as a factor that could influence gastrointestinal or respiratory outcomes. In the current study, we found that patients who were receiving higher proportions of thin liquids from formula or added water were more likely to undergo chest x-rays in the year after starting a commercial blenderized formula than those receiving lower proportions of thin liquids. This finding remained significant even when accounting for underlying pulmonary comorbidities and aspiration, suggesting that the addition of free water to blends may independently impact respiratory outcomes.
Blenderized tube feeds are becoming increasingly common as a desired feeding option in patients with enteral tubes.1,13 Research in this area is growing, and studies have shown multiple benefits, including reduced gastrointestinal and respiratory symptoms, increased parental satisfaction, and reduced costs.2–7,14 The pathophysiology underlying the improved clinical outcomes associated with blenderized formulas is still being studied, but a main hypothesized mechanism relates to blends being higher viscosity than standard formulas, which promotes movement of feeds from the gastric body to the antrum and may mediate reflux-associated respiratory symptoms.8,10–12,15 Importantly, this means that details surrounding blend preparation and administration can significantly change the blend and impact patient outcomes. A prior study from our group demonstrated that the addition of water to blends can drastically reduce blend viscosity.9 For example, adding just 30mL of water to a 240mL pouch of Real Foods Blends decreased viscosity by 50%. The results of the current study suggest that, in addition to the reduction in viscosity seen ex vivo, the addition of extra thin liquids into a patient’s feeding regimens may mitigate some of the positive effects typically associated with blenderized feeds.
Given that water is typically given in an effort to maintain patient hydration, a desire for thicker feeds must be balanced with the need for adequate amounts of total fluid. Fluid needs may be estimated using equations, such as the Holliday-Segar method, which was developed to calculate parenteral fluid needs in hospitalized patients.16 The practical application of this type of calculation to tube fed patients of varying ages and comorbidities has not been studied. In the current study, regardless of the viscosity of the commercial blend being received, there was variability in the amount of added water and the percent of total estimated fluid received. As a measure of hydration status, one can record urine specific gravity values when patients were well; however, only two patients had results recorded during the follow-up period limiting interpretation of these data. We also assessed serum chemistries, and there was no statistically significant difference in the frequency of abnormally chemistries between groups, suggesting that receiving minimal thin liquids did not result in electrolyte derangements. Prospective studies would be useful to better assess hydration needs in medically complex pediatric patients on tube feeds, as well as to evaluate non-thin options, like purees, for meeting those needs. In addition to hydration, providers and families sometimes worry about fluids in the context of constipation, which was present in 62% of our patient sample. While we did not evaluate changes in constipation as an outcome in this study, the ESPGHAN/NASPGHAN guideline for treatment of functional constipation recommends against increasing fluids for treatment of constipation, as there is no high quality evidence to support this practice.17 Further research on the relation between water intake, blend administation and constipation in medically complex patients are needed to determine if there is any impact of water on stool frequency or consistency.
There are a few limitations to this study. The main limitation was the use of retrospective chart review, which limited the level of detail that could be obtained. While chest x-rays were an objective and available proxy for respiratory-related events, prospective data on respiratory and other outcomes would be most useful to better define the relationship between added thin liquids and respiratory outcomes. Another limitation was the heterogeneity of the patient sample, which impacts the generalizability of the findings, however we tried to take this into account by including in our multivariable analysis relevant comorbidities and aspects of the feeding regimen, including whether the patient was receiving any oral feeds and whether the patient was receiving bolus or continuous feeds. A final limitation was the inability to fully assess patient hydration given that so few patients had urine specific gravity values. This is an important factor to consider when thinking about the amount of water added to blends, and it merits further study.
In summary, this study demonstrated wide variability in care practices related to free water administration in patients receiving blenderized feeds. Chest x-rays were less common in children receiving minimal thin liquids. The addition of extra water or other thin liquids to tube feeds should be considered as a factor that may modify pulmonary outcomes.
What is known
Blenderized feeds reduce gastrointestinal and respiratory symptoms and improve quality of life.
The high viscosity of blends may reduce reflux and reflux-associated symptoms.
The addition of extra water to blended formulas reduces feed viscosity.
What is new
The amount of free water given to patients on blenderized feeds varies widely, with some patients receiving up to 1 liter of water in addition to tube feeds.
Patients receiving minimal thin liquids were less likely to undergo chest x-rays in the year after blend initiation.
The addition of extra water to tube feeds should be considered as a factor that may modify patient outcomes.
Funding Sources:
Supported by National Institutes of Health R01 DK097112.
List of abbreviations
- GI
gastrointestinal
- IDDSI
International Dysphagia Diet Stardardization Initiative
Footnotes
Conflicts of Interest: The authors have no conflicts of interest to disclose.
References
- 1.Chandrasekar N, Dehlsen K, Leach ST, Krishnan U. Exploring Clinical Outcomes and Feasibility of Blended Tube Feeds in Children. J Parenter Enter Nutr. Published online February 17, 2021:jpen.2062. [DOI] [PubMed] [Google Scholar]
- 2.Pentiuk S, O’Flaherty T, Santoro K, Willging P, Kaul A. Pureed by Gastrostomy Tube Diet Improves Gagging and Retching in Children With Fundoplication. J Parenter Enter Nutr. 2011;35:375–379. [DOI] [PubMed] [Google Scholar]
- 3.Gallagher K, Flint A, Mouzaki M, et al. Blenderized Enteral Nutrition Diet Study: Feasibility, Clinical, and Microbiome Outcomes of Providing Blenderized Feeds Through a Gastric Tube in a Medically Complex Pediatric Population. J Parenter Enter Nutr. 2018;42:1046–1060. [DOI] [PubMed] [Google Scholar]
- 4.Hron B, Fishman E, Lurie M, et al. Health Outcomes and Quality of Life Indices of Children Receiving Blenderized Feeds via Enteral Tube. J Pediatr. 2019;211:139–145.e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Batsis ID, Davis L, Prichett L, et al. Efficacy and Tolerance of Blended Diets in Children Receiving Gastrostomy Feeds. Nutr Clin Pract. 2020;35:282–288. [DOI] [PubMed] [Google Scholar]
- 6.Trollip A, Lindeback R, Banerjee K. Parental Perspectives on Blenderized Tube Feeds for Children Requiring Supplemental Nutrition. Nutr Clin Pract. 2020;35:471–478. [DOI] [PubMed] [Google Scholar]
- 7.Kernizan D, Mintz D, Colin M, et al. Outcomes and Safety of Blenderized Tube Feedings in Pediatric Patients: A Single Center’s Experience. J Pediatr Gastroenterol Nutr. 2020;71:e124–e128. [DOI] [PubMed] [Google Scholar]
- 8.Sullivan Mary, Pearl Sorreda-Esguerra, Platon MB, et al. Nutritional analysis of blenderized enteral diets in the Philippines. Asia Pac J Clin Nutr. 2004;13:385–390. [PubMed] [Google Scholar]
- 9.Hron B, Rosen R. Viscosity of Commercial Food-based Formulas and Home-prepared Blenderized Feeds. J Pediatr Gastroenterol Nutr. 2020;70:e124–e128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wenzl TG, Schneider S, Scheele F, Silny J, Heimann G, Skopnik H. Effects of Thickened Feeding on Gastroesophageal Reflux in Infants: A Placebo-Controlled Crossover Study Using Intraluminal Impedance. Pediatrics. 2003;111:e355–e359. [DOI] [PubMed] [Google Scholar]
- 11.Nishiwaki S, Araki H, Shirakami Y, et al. Inhibition of Gastroesophageal Reflux by Semi-solid Nutrients in Patients With Percutaneous Endoscopic Gastrostomy. J Parenter Enter Nutr. 2009;33(5):513–519. [DOI] [PubMed] [Google Scholar]
- 12.Adachi Kyoichi Shizuku Toshitsugu. Efficacy of half-solid nutrient for the elderly patients with percutaneous endoscopic gastrostomy. J Clin Biochem Nutr. 2011;48:226–229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Bennett K, Hjelmgren B, Piazza J. Blenderized Tube Feeding: Health Outcomes and Review of Homemade and Commercially Prepared Products. Nutr Clin Pract. 2020;35:417–431. [DOI] [PubMed] [Google Scholar]
- 14.Bobo E Reemergence of Blenderized Tube Feedings: Exploring the Evidence. Nutr Clin Pract. 2016;31(6):730–735. [DOI] [PubMed] [Google Scholar]
- 15.Kanie J, Suzuki Y, Akatsu H, Kuzuya M, Iguchi A. Prevention of Late Complications by Half-Solid Enteral Nutrients in Percutaneous Endoscopic Gastrostomy Tube Feeding. Gerontology. 2004;50(6):417–419. [DOI] [PubMed] [Google Scholar]
- 16.Holliday M, Segar W The Maintenance Need for Water in Parenteral Therapy. Pediatrics. 19(923):12. [PubMed] [Google Scholar]
- 17.Tabbers MM, DiLorenzo C, Berger MY, et al. Evaluation and Treatment of Functional Constipation in Infants and Children: Evidence-Based Recommendations From ESPGHAN and NASPGHAN. J Pediatr Gastroenterol Nutr. 2014;58(2):258–274. [DOI] [PubMed] [Google Scholar]