Dietary caloric density and osmolality influence gastroesophageal reflux in infants

GASTROENTEROLOGY1989:97:8014

Dietary Caloric Density and Osmolality Influence Gastroesophageal Reflux in Infants JAMES Department Virginia

L. SUTPHEN of Pediatrics,

and VIVIAN L. DILLARD Children’s

Medical

In a group of 19 infants being evaluated for gastroesophageal reflux, we investigated the effects of various carbohydrate solutions (glucose polymers, 5% dextrose in water, and 10% dextrose in water) on the rate of postcibal gastroesophageal reflux during the first 2 h after a test feeding. The highosmolality feeding (10% dextrose in water) produced significantly more postcibal gastroesophageal reflux over the entire 2-h interval. The major difference occurred in the second postcibal hour when the amount of gastroesophageal reflux was persistently high for 10% dextrose in water in contrast to the other feedings. We speculate that more rapid gastric emptying of low-osmolality solutions may account for these differences. Clear liquid feeding composition should be standardized during pH testing. Low-osmolality glucose polymer solutions may be more easily tolerated by infants with gastroesophageal reflux who require carbohydrate or fluid supplements.

G

astroesophageal reflux (GER) in infants is now recognized to be a common diagnostic entity. An accepted technique for the diagnosis of GER in infants involves intraesophageal pH monitoring. During esophageal pH studies clear liquid feedings [generally sugar-water or fruit juice) are often utilized to facilitate the detection of postcibal reflux as these feedings do not buffer gastric acid as much as mixed feedings (1). The osmolality and caloric density of clear liquid feedings and, in fact, routine infant feedings may vary considerably. Therefore, the effects of these dietary factors on rates of postcibal GER may be potentially important variables that could influence diagnosis during postcibal testing (pH probe, barium swallow, isotope scans) and potential therapy with dietary modification. It is the purpose of this report to document the effects of osmolality and caloric density of carbohydrate solu-

Center,

University

of Virginia,

Charlottesville,

tions on the rate of early postcibal gastroesophageal reflux as detected on esophageal pH testing.

Materials and Methods Nineteen infants referred to the University of Virginia Medical Center for evaluation of esophageal reflux during an interval from April through October 1986 were candidates for this study. Descriptive data concerning the group are outlined in Table 1. The reasons for evaluation included suspected apnea (n = 12), recurrent emesis (n = 8), and pulmonary symptoms (n = 15). The pH probe (model MI-506; Microelectrodes, Londonderry, N.H.) was attached to a Radiometer PHMGI pH meter (Radiometer America, Westlake, Ohio) and standardized with pH 4 and 7 buffer solutions just before the examination. It was passed through the nares into the stomach where an acid pH of <4 was verified. It was then withdrawn to a point previously calculated from a nomogram to represent 87% of the distance from the nares to the lower esophageal sphincter (2). Its position was verified by chest radiograph to be in the distal third of the esophagus above the diaphragm. The pH was continuously monitored on a Hewlett Packard graphic recorder (Hewlett-Packard, Palo Alto, Calif.). After a 4-h fast, each infant was given three clear liquid carbohydrate feedings (15 ml/kg body wt) in random sequence. The feedings were 5% dextrose in water (D5W), 20 kcal/dl, 297 mosmol/kg water; 10% dextrose in water (DlOW), 40 kcal/dl, 594 mosmollkg water; and glucose polymers (Polycose; Ross Laboratories, Columbus, Ohio), 40 kcalidl, 180 mosmolikg water. The dextrose solutions were standard commercial enteral solutions prepared by Ross Laboratories. The Polycose solution was prepared from a weighed quantity (10.5 g) of Polycose powder diluted to 100 ml with sterile water (osmolality and caloric

used in this paper: D5W, 5% dextrose in water; dextrose in water; EAE, esophagealacidexposure; GER, gastroesophageal reflux. 0 1989 by the American Gastroenterological Association 0018~5085/89/$3.50 Abbreviations

DlOW, 10%

602

SUTPHEN AND DILLARD

Table

1. Descriptive

GASTROENTEROLOGY Vol.

Data” Mean

Postnatal age (mo) Weight (kg) Length (cm) Minutes of EAElh (overnight study] Duration of overnight study (h)

Median

3.7

2.5

0.7-13.2

5.1

1.8-7.5

58.0

59.3

4.5

3.9

0.2-17.0

16.5

12.8

8.5-58.5

EAE, esophageal acid exposure. “Nineteen preterm; 10 female, 9 male.

infants:

hours analyzed separately. For each feeding the amounts of GER during the first and second postcibal hour were compared by paired t-test. All statistical analyses were performed using National Institutes of Health Clinfo software. This study was approved by the University of Virginia Human Investigation Committee (May 15, 1985).

Range

4.8

45-69

11

Results 8

term,

density figures are taken from Ross Laboratories product information). The pH of the three solutions was weakly acid (pH 5-5.5). Immediately following each feeding the infant was burped and placed in a horizontal prone position and a 2-h pH monitoring interval ensued. If the infant objected to this position, the position of preference was used and repeated after each feeding for that infant. The interval was divided into an early (first 60 min) and late [second 60 min) phase. No positional measures or other stresses were employed to induce or control GER. None of the infants objected to the prone position; however, the oldest 4 infants (one IO-mo-old and three 5-mo-old infants) moved and rolled in their crib without restrictions during the postcibal period. Enforcing a uniform position in these infants required restraints and led to irritability and crying. The minimum interval between feedings was 2 h. Although we attempted to administer all three feedings to each infant, for clinical reasons [time constraints) unrelated to feeding intolerance, 3 infants did not receive Polycose, 2 did not receive DlOW, and 1 did not receive D5W. Because of recorder malfunction, 1 infant was monitored for only 1 h after the D5W and Polycose feedings and 2 infants were monitored for only 1 h after the DlOW feedings. After completion of the clear liquid feedings, the pH probe was left in place overnight and the pH was continuously recorded while the infant was on a demand schedule with regular infant formula only (Similac, Ross Laboratories). After the study the probe was removed and restandardized with agreement to +-0.5 pH units. A period of reflux was defined as an episode of esophageal pH of ~4. All episodes >l2 s in duration were counted and expressed as total minutes of esophageal acid exposure (EAE). The overnight pH probe data were analyzed separately and reported as minutes of EAE per hour of recording. The effects of caloric density and osmolality were evaluated by analysis of variance, followed (when significant) by Dunnett’s test of multiple comparison with DlOW as the control and the first and second postcibal

Table Feed D5W

2. Minutes

of Esophageal n 18

DlOW

17

Polycose

16

Acid

97, No. 3

Summary

data describing

the amount

of EAE

over the two postcibal hours are presented in Table 2. During the first postcibal hour the amount of EAE did not differ between the three feedings by analysis of variance. At the 0.05 level of significance the power of this analysis to detect a 7-min difference is 50%~80%. During the second postcibal hour and the total 2-h period, both D5W and Polycose produced significantly less EAE than DlOW (p < 0.05 by analysis of variance followed by Dunnett’s test). There was a decrease in the amount of EAE in the second versus the first postcibal hour for all feedings, but the change was statistically significant for D5W and glucose polymers only (Table 2). At the 0.05 level of significance the power of this analysis to detect a 7-min difference is ~50%. The rate of EAE observed during either the early or late postcibal phases of all feedings did not correlate with the amount seen on prolonged pH probe recordings with regular infant formula by linear regression analysis. (Probability values for these regressions ranged from 0.10 to 0.94.) The 9 patients with the least amount of reflux on prolonged pH probe testing (1.4 + 1.1; range, 0.2-2.9 min/h) demonstrated a similar difference in GER between carbohydrate solutions and early and late postcibal periods. However, the groups were too small to demonstrate statistical significance. Discussion It has been suggested that slow rates of gastric emptying are associated with an increase in GER (3). Others have suggested that there is no relationship between gastric emptying and GER (4). To date, investigations have compared gastric emptying between people with and without GER. Studies have not been directed toward the question of response of

Exposure

First postcibal hour 9.5 t 9.7 (a] 16.7 2 14.8(d) 9.7 2 5.9

(g]

n 17 15 15

Second postcibal hour 2.5 2

4.3(b)

11.9 2 16.0(e] 2.9 t 5.5 (h)

n

Total

(2

h)

17

12.0 * 11.3(c)

15

28.6

15

12.6 f 8.0(i)

D5W, 5% dextrose in water: DIOW, 10% dextrose in water. Values are mean 2 SD. Paired t-tests: a/b, p = 0.008; d/e, p = 0.004. Analysis of variance: a/d/g, NS; b < e, h < e, p < 0.05 by Dunnett’s test. c < f, i < f, p < 0.05 by Dunnett’s test.

0.15;

2 28.4 (f]

g/h, p =

September 1989

DIET INFLUENCES GER IN INFANTS

Table 3. Total Sugar Concentration and Osmolality of Selected Fluids Sugar Apple juice Grape juice Kool Aid 7-up Orange juice D5W DlOW Polycose 10%

(g/d])”

Osmolality (mosmol/kg H,O)”

10.7-12 15.1 10.4 10 10.9 5 10 10

654-734 1167-1190 407-594 523-548 553-672 297 594 180

D5W, 5% dextrose in water; DlOW, 10% dextrose a References 14 and 15. b Reference 16.

in water.

individuals with GER to variations in diet that influence gastric emptying. Gastric distention is positively correlated with rates of postcibal GER (5,6) expressed as esophageal acid exposure. Therefore, if gastric emptying were delayed and gastric distention prolonged, the amount of postcibal GER would be expected to increase. Osmolality and caloric density of liquid feedings are known to influence the rate of gastric emptying (7-12). It has been shown that high-osmolality solutions empty more slowly from the stomach than isotonic solutions (8-12). The rate of gastric emptying of glucose solutions decreases as the caloric density and osmolality of the solution increase (10,ll). Isocaloric solutions of glucose polymers empty from the stomach more rapidly and induce less gastric secretion than glucose monomers (8,ll). By inference, these dietary factors could affect rates of postcibal GER as well. Table 3 presents the carbohydrate concentration and osmolality of some common fluids often given to infants as supplements or therapy for recurrent emesis (14-16). Our data suggest that when caloric density is increased by increasing osmolality of carbohydrate solutions, there is an increase in postcibal reflux. However, when osmolality is not increased, there is less effect. These findings might be altered by adjusting the feeding volume, consistency, or source of calories (fat, carbohydrate, or protein). We chose a relatively small volume (15 ml/kg body wt) to ensure that enteral tube feedings would not be necessary to achieve identical volumes of the different feedings. It is interesting that the amount of postcibal GER did not correlate with the amount seen on prolonged esophageal monitoring (overnight) on formula diet in infants. This is in part due to the buffering effect of the formula diet on postcibal gastric pH, which obscures the detection of postcibal reflux. Therefore, the formula feeding data on prolonged pH monitoring are more indicative of late postcibal GER. Late unbuffered postcibal GER may be more harmful as a

803

cause of esophagitis (13). However, many clinical symptoms of GER are noted by parents to occur in the early postcibal period (e.g., crying, choking, apnea, spitting) and may be associated with a decrease in esophageal pH during testing. Aspiration of stomach contents is potentially a greater problem during the early postcibal period when the stomach is full. Therefore, careful examination of the postcibal period for GER is a valuable component of the diagnostic evaluation of infants. In fact, nuclear medicine and barium radiographic studies generally assess only postcibal GER. Although much has been written about the effects of diet on diarrhea in infants, there is no information to date on the effects of dietary constituents on GER. Our data suggest that the contents of clear liquid feedings should be standardized during esophageal pH probe testing. When clear liquids are given to children who have GER it is reasonable to avoid those with simple sugars that have high osmolalities (e.g., fruit juices). Diagnostic tests that specifically examine postcibal GER (barium studies and nuclear medicine scans) should be standardized in terms of caloric density, osmolality, and volume of the test solution.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10. 11.

Sutphen JL, Dillard VL. Effects of maturation, gastric acidity, and symptom complex on gastroesophageal reflux. Am J Dis Child 1986;140:1062-5, Strobe1 CT, Byrne WJ, Ament ME, Euler AR. Correlation of esophageal lengths in children with height: application to the Tuttle test without prior esophageal manometry. J Pediatr 1979:94:81-4. Hillemeier AC, Grill BB, McCallum R, Gryboski J. Esophageal and gastric motor abnormalities in gastroesophageal reflux during infancy. Gastroenterology 1983;84:741-6. Rosen PR, Treves S. The relationship of gastroesophageal reflux and gastric emptying in infants and children: concise communication. J Nucl Med 1984;25:571-4. Holloway RH, Hongo M, Berger K. McCallum RW. Gastric distention: a mechanism for postprandial gastroesophageal reflux. Gastroenterology 1985;89:779-84. Sutphen JL. Dillard VL. Effect of feeding volume on gastroesophageal reflux in infants. J Pediatr Gastroenterol Nutr 1988:7:185-8. Siegel M, Lebenthal E, Krantz B. Effect of caloric density on gastric emptying in premature infants. J Pediatr 1984;104: 118-22. Costalos C, Russell G, Al Rahim Q, Blumenthal I, Hanlin S, Ross I. Gastric emptying of Caloreen meals in the newborn. Arch Dis Child 1980;55:883-5. Minami H, McCallum RW. The physiology and pathophysiology of gastric emptying in humans. Gastroenterology 1984; 86:1592-610. Brener W, Hendrix TR, McHugh PR. Regulation of the gastric emptying of glucose. Gastroenterology 1983:85:76-82. Foster C, Costill DL, Fink WJ. Gastric emptying characteristics of glucose and glucose polymer solutions. Res Q Exercise Sports 1980:51:299-305.

604

SUTPHEN AND DILLARD

12. Meeroff JC, Go VLW, Phillips SF. Control of gastric emptying by osmolality of duodenal contents in man. Gastroenterology 1975;68:1144-51. 13. Hyams JS, Ricci A, Leichtner AM. Clinical and laboratory correlates of esophagitis in young children. J Pediatr Gastroenter01 Nutr 1988;7:52-6. 14. Hurst WJ, Martin RA, Zoumas BL. Application of HPLC to characterization of individual carbohydrates in foods. J Food Sci 1979;44:892-5. 15. Pennington JA, Church HN. Bowe’s and Church’s food values of portions commonly used. 13th ed. Philadelphia: JB Lippincott, 1980. 16. Wendland BE, Arbus GS. Oral fluid therapy: sodium and

GASTROENTEROLOGY Vol. 97, No. 3

potassium content and osmolality of some commercial “clear” soups, juices and beverages. Can Med Assoc J 1979; 121:564-71.

Received October 6, 1988. Accepted March 27, 1989. Address requests for reprints to: James L. Sutphen, M.D., The University of Virginia, Children’s Medical Center, Department of Pediatrics, Box 386, University of Virginia Medical Center, Charlottesville, Virginia 22908. This work was supported in part by General Clinical Research Center grant MOlRR00847 from the National Institutes of Health.