The infant respiratory chamber. Note the glove ports surrounding the chamber, allowing for unrestricted access.
Continuous measurements of the index of physical activity. Data points represent the mean of five 1-minute summaries for all infants. Data shown are grouped according to carbohydrate absorption (top) and type of juice (bottom). Physical activity was higher (P<.001) in infants with incomplete carbohydrate absorption (breath hydrogen level ≥20 ppm above baseline). When grouped according to juice (pear vs white grape), physical activity was higher (P<.01) in infants fed pear juice.
Continuous measurements of metabolic rate. Data points represent the mean of five 1-minute summaries of all infants. Data shown are grouped according to carbohydrate absorption. Energy expenditure was higher (P<.05) in infants with incomplete carbohydrate absorption (breath hydrogen level ≥20 ppm above baseline).
Continuous measurements of the respiratory quotient. Data points represent the mean of five 1-minute summaries of all infants. Data shown are grouped according to carbohydrate absorption. Respiratory quotients were higher (P<.05) in infants with incomplete carbohydrate absorption (breath hydrogen level ≥20 ppm above baseline).
Customize your JAMA Network experience by selecting one or more topics from the list below.
Cole CR, Rising R, Lifshitz F. Consequences of Incomplete Carbohydrate Absorption From Fruit Juice Consumption in Infants. Arch Pediatr Adolesc Med. 1999;153(10):1098–1102. doi:10.1001/archpedi.153.10.1098
Most infants consume fruit juices by 6 months of age. However, fruit juices containing sorbitol may be associated with carbohydrate malabsorption without clinical symptoms. We hypothesized that increased physical activity and metabolic rate may be associated with carbohydrate malabsorption.
Physical activity and metabolic rate were determined in 14 healthy infants ([mean±SD] age, 5.1±0.8 months; weight, 7.8±1.1 kg; length, 67±4.2 cm; and body fat, 26%±5%) for 3 hours in a respiratory chamber. Seven were fed pear juice, and the other 7 were fed white grape juice (120 mL) after a 2-hour fast. Pear juice contains sorbitol and a high fructose–glucose ratio, whereas white grape juice is sorbitol free and has a low fructose–glucose ratio. Carbohydrate absorption was determined by breath hydrogen gas analysis. The study was double-blinded.
When compared with the infants without carbohydrate malabsorption (peak breath hydrogen level <20 ppm above baseline), 5 of the 7 infants fed pear juice and 2 of the 7 infants fed white grape juice exhibited carbohydrate malabsorption (peak breath hydrogen level ≥20 ppm above baseline; P<.01). These infants also exhibited both increased physical activity (P<.001) and metabolic rate (P<.05) after juice consumption in comparison with infants with normal carbohydrate absorption. When grouped according to the type of juice consumed, only infants fed pear juice exhibited increases in physical activity (P<.01).
Carbohydrate malabsorption is associated with increased physical activity and metabolic rate in infants. Most of the infants who had carbohydrate malabsorption consumed pear juice. Therefore, fruit juices containing sorbitol and high levels of fructose may not be optimal for young infants.
FRUIT JUICES are an integral part of an infant's diet because of taste, availability, low price, and positive exposure as a healthy snack food. This has resulted in a multimillion dollar market for fruit juices packaged exclusively for infant consumption.1,2 Intake usually starts when supplemental foods are introduced to the infant at 4 to 6 months of age.3,4 According to a survey conducted by juice manufacturers, more than 90% of all infants consume some type of fruit juice by 1 year of age.4
The chemical makeup of such a popular food and its absorption from the infant's gut have been the interest of several investigators.4-7 For example, juices containing sorbitol and high fructose-to-glucose ratios, such as apple or pear juice, exhibit incomplete carbohydrate absorption in children from 6 months to 5 years of age.4,5,7,8 In another study, children younger than 6 years incompletely absorbed oral fructose when it was administered alone (0.7 to 2.0 g/kg); if given with equimolar amounts of glucose, its absorption was enhanced.7 These findings have resulted in an advisory from the American Academy of Pediatrics Committee on Nutrition for moderation in the amount of sorbitol-containing fruit juices fed to children.9
The malabsorption of carbohydrates produces excess hydrogen gas. This has been suggested as a cause of minor irritability or colic.10,11 We hypothesize that increased physical activity and metabolic rate are the result of carbohydrate malabsorption after the consumption of sorbitol-containing fruit juices. We used a recently developed infant respiratory chamber12 to simultaneously evaluate carbohydrate absorption from fruit juice consumption in relation to physical activity and metabolic rate.13 These tests were performed under double-blinded conditions on infants who had ingested either pear or white grape juice.
Fourteen healthy infants were recruited from the pediatric ambulatory clinics of Maimonides Medical Center, Brooklyn, NY (Table 1). Children with existing medical problems, antibiotic usage within the previous month, diarrhea, or other gastrointestinal tract symptoms were excluded from the study. Informed consent was obtained from at least 1 parent of each child. This study was approved by the Research Committee of Maimonides Medical Center.
The infants entered the protocol after a 2-hour fast. On the day of the visit, parents completed a questionnaire regarding the infant's health status, dietary intake, and juice consumption. Anthropometric measurements of weight, supine length, and skin-fold thickness at the subscapular, triceps, and flank sites were made by the same investigator (C.R.C.). Body weights were the average of 2 measurements obtained with an infant scale (Detecto Scales Inc, Brooklyn, NY). Supine lengths (crown to heel) were measured in duplicate with a horizontal stadiometer (Perspective Enterprises, Kalamazoo, Mich). Skin-fold thicknesses were the mean of 3 measurements on the right side of the body made by means of a skin-fold caliper (Lange; Beta Technology, Cambridge, Md) according to the standard procedure.14 Body fat and fat-free mass were calculated by appropriate equations.15
White grape and pear juices were selected for the study because of their constituent carbohydrates. White grape juice contains equal amounts of glucose and fructose and no sorbitol, whereas pear juice contains approximately a 3:1 ratio of fructose to glucose along with sorbitol. The carbohydrate content of pear juice is similar to that of apple juice. Pear juice was the sorbitol-containing juice that was provided to us at the time of the study. The juices were provided by Welch Foods Inc, Concord, Mass, and were coded. Neither the investigators nor the parents knew which juice the infant was receiving. There were no detectable differences in regard to appearance and odor between the 2 types of juice.
All juices were fed at room temperature in 120-mL servings, which is typical for children of these ages.4 The amounts of sorbitol, sucrose, fructose, and glucose contained in 120 mL of each juice fed in this study are as follows: white grape juice, 0, 0, 8.8, and 8.4 g, respectively; and pear juice, 2.4, 1.1, 7.5, and 2.7 g, respectively. Eleven infants consumed the 120 mL, whereas 3 infants consumed only 90 mL. Since the infants had fasted for 2 hours before juice feeding, they consumed the juice within a couple of minutes. None of the infants vomited during the study. No other food or beverage was permitted.
Physical activity and metabolic rate of the infants were measured by indirect calorimetry by means of a new infant respiratory chamber.12,13 The index of physical activity for each infant was calculated by measuring the oscillations in weight generated by the infant on a specially designed platform within the respiratory chamber (Figure 1). The index of physical activity was the sum of 1-minute oscillations in weight (in grams) divided by the infant's body weight (in kilograms). Metabolic rate (kilocalories per minute) was continuously calculated according to the method of Ravussin et al.16 Physical activity and metabolic rate data were summarized every 5 minutes by the computer software. The infants were monitored by 1 investigator (C.R.C.) for the entire testing period, and an activity log sheet was kept. This included the amount of time infants cried during the procedure. Basal metabolic rate (kilocalories per minute) was equal to the sum of all metabolic rate summaries calculated while the infant was awake, in a postabsorptive state (at least 1 hour after the last meal, depending on the feeding routine of the infant), and with a minimum amount of movement. The respiratory chamber and method of physical activity measurement were validated as described previously12; however, the following is a brief description of the infant respiratory chamber.
A small Plexiglas enclosure measuring 1 m3 was part of the infant respiratory chamber (Figure 1). A portable instrument rack housed the oxygen and carbon dioxide analyzers; flowmeter; barometric pressure, temperature, and humidity sensors; electric eye controller; and computer equipment. There is enough room in the chamber enclosure for storage of diapers and other supplies necessary for the care of the infant and breath hydrogen gas (BH2) measurements. One unique feature incorporated into the chamber enclosure is 3 pairs of long rubber gloves that were installed around the entire enclosure (Figure 1). These gloves allowed the parents unrestricted access to the infant and were also used by the investigators for BH2 sampling without corrupting the environment within the chamber. The volume occupied by the gloves during use was corrected for with the aid of an electric eye (Photoswitch 42GRR/GRL-9000, Allen-Bradley, Santo Domingo, Dominican Republic) whenever its beam was broken by the gloves. The volume occupied by the gloves was calculated to be 6 L.
Room air was pulled through the chamber and exhausted to the outside via a fan mounted in the exhaust line of the chamber enclosure. Oxygen and carbon dioxide concentrations, flow rate, chamber temperature, barometric pressure, and humidity were measured continuously on the exhaust side of the system. Oxygen and carbon dioxide concentrations were determined by means of 2-channel differential analyzers (Hartmann & Braun, Frankfurt, Germany). One channel served as the reference, using room air, while the other measured oxygen and carbon dioxide concentrations within the chamber. This provided continuous corrections for changes in ambient oxygen and carbon dioxide concentrations. Flow rate was converted to standard temperature, pressure dry conditions by measuring barometric pressure (millimeters of mercury), temperature (degrees Celsius), and humidity (percentage) within the exhaust line of the system. A sample gas cooler removed the moisture from the air sample before analysis by the oxygen and carbon dioxide analyzers.
After 30 minutes of acclimation within the chamber, breath samples were obtained from the infant immediately before the juice was fed. Thereafter, expired air was sampled at 30-minute intervals for 2 hours. Breath samples were collected by means of kits (Pediatric Breath Collection Kits) modified from the system for children (GaSampler; QuinTron Instrument Co, Milwaukee, Wis).17 The modifications were as follows: (1) the mouthpiece was replaced with a soft, circular, infant-sized, 6.0-cm face mask (Vital Signs Inc,Totowa, NJ) and (2) a nonrebreathing valve was inserted into the T-port.18 The face mask was held over the subject's mouth and nose, and the collection bag was filled while the infant breathed normally. Collected expired airwas analyzed for BH2 within 3 hours of collection by means of a microlyzer (Model SC; QuinTron Instrument Co). The infant was released to the family, who were instructed to report any gastrointestinal tract symptoms that developed in the infants within 48 hours.
The following BH2-related variables were quantified10: (1) peak BH2, expressed in parts per million of hydrogen; this was the highest BH2 level attained by each subject after juice intake; (2) area under the curve, which is the area below all the measured BH2 levels, calculated by the summation of each one of the segmental areas at each of the 30-minute intervals; and (3) incomplete carbohydrate absorption, defined as a rise of BH2 20 ppm or greater above the value obtained at the end of the acclimation period or baseline.
Before statistical analysis, we disclosed the code of the juices to determine which infants received pear juice and which received white grape juice. Data were analyzed with SPSS software (SPSS Inc, Chicago, Ill). Data were grouped either by juice (white grape vs pear) or by carbohydrate absorption status (peak BH2 <20 ppm vs ≥20 ppm in comparison with baseline). Differences in physical characteristics and BH2 analysis between groups were determined by independent t test. The computer software for the infant respiratory chamber summarizes metabolic rate data every 5 minutes. There are 30 five-minute summary periods for physical activity and metabolic rate for each infant tested. Juice was fed starting after 30 minutes (during summary period 7). The first 6 summary periods were considered baseline, during which time the infant was adapting to the environment within the chamber, while summary periods 7 though 30 were considered the observation periods after juice feeding. Expired air for BH2 analysis was collected during summary periods 7, 13, 19, and 25 and at the conclusion of the metabolic rate measurement. For each infant, any summary period in which there was outside interaction by parents or staff was eliminated before data analysis. This included the 5 summaries in which expired air was collected for BH2 analysis. Differences between groups during the observation period were determined by analysis of variance. Significance was established at the 5% level of probability (P<.05).
Analysis of the initial questionnaire completed by the parents at the beginning of the study showed that apple juice was the juice of choice in 43% of the families. The other infants consumed combinations of apple-pear and grape juice, except for 1 infant who never drank fruit juice before the study (who was fed grape juice during the study). The mean±SD juice intake was 114±66 mL/d. This was independent of the type of juice and carbohydrate absorption status.
No anthropometric differences in the infants were detected regardless of carbohydrate absorption status or the type of juice fed (Table 1). All infants were growing normally between the 5th and 95th percentiles on appropriate sex-specific National Center for Health Statistics growth charts.19 The weight-for-height ratios were appropriate in all patients.
Two infants fed white grape juice and 5 infants fed pear juice had incomplete carbohydrate absorption (Table 2). These infants had peak BH2 measurements ranging from 25 to 67 ppm above baseline. Five infants fed white grape juice and 2 infants fed pear juice fully absorbed carbohydrates. These infants had peak BH2 measurements ranging from 3 to 19 ppm above baseline. Those with carbohydrate malabsorption had significantly higher peak BH2 excretion along with a significantly greater area under the curve. The mean level of BH2 excreted in the carbohydrate malabsorption group was more than 4 times that measured for infants with no carbohydrate malabsorption (Table 2). When the experiment was started, 3 of the 14 infants consumed only 90 mL of juice. Two of them were fed white grape juice and had complete carbohydrate absorption (peak BH2, 9.0 and 5.0 ppm above baseline, respectively), while 1 was fed pear juice and had incomplete carbohydrate absorption (peak BH2, 34 ppm above baseline).
Infants with incomplete carbohydrate absorption had significantly increased physical activity and a greater metabolic rate after juice consumption in comparison with infants who fully absorbed carbohydrates (Figure 2 and Figure 3). Infants with incomplete carbohydrate absorption also had significantly greater respiratory quotient after juice consumption in comparison with infants who fully absorbed carbohydrates (Figure 4). When analyzed according to the type of juice fed, physical activity was significantly increased with pear juice consumption (Figure 2).
Regardless of how the data were analyzed, no differences were found for the amount of parental interaction, basal metabolic rate, or the percentage of crying during the observation period (Table 2). There were no reports of gastrointestinal tract symptoms up to 48 hours after the study in any of these infants.
Previous studies have shown a relationship between exhaled BH2 response and incomplete carbohydrate absorption in infants and children after consuming fruit juices.4-6,8 However, no studies have determined any specific clinical effects that may be related to incomplete carbohydrate absorption after fruit juice ingestion in infants. We used a new infant respiratory chamber that determined physical activity and metabolic rate in relation to carbohydrate absorption.12,13 We found much greater physical activity, metabolic rate, and respiratory quotients in infants with incomplete carbohydrate absorption after consuming fruit juice. It appears that most of the increase in physical activity and metabolic rate occurred during the last hour of the study. Most of the infants with incomplete carbohydrate absorption consumed pear juice. These results suggest that fruit juices containing sorbitol, and with high fructose-to-glucose ratios, are not well absorbed by most 6-month-old infants, and this is associated with consistent changes in physical activity and metabolic rate.
Our results showed a significant increase in exhaled hydrogen gas concomitant with gastrointestinal tract malabsorption. It is feasible that gastrointestinal tract discomfort, expressed as an increase in physical activity and metabolic rate, could also be the result of gas that is produced while unabsorbed sorbitol and fructose are fermented in the lower part of the colon.8 However, this discomfort may not be severe enough to initiate any additional crying, but it is enough to cause more activity and increase the metabolic rate. Although the amount of gas formed by unabsorbed carbohydrates from juice intake in infants has not been quantified, it has been found in adults and children that approximately 15% of the hydrogen gas produced by colonic fermentation from unabsorbed carbohydrates is expelled though the lungs, while the rest is expelled as flatulence.20 Lactose-intolerant individuals produce about 50 mL of gas per 29 mL of cow's milk consumed. Cow's milk contains 4.6% lactose, or 1.4 g per 29 mL.21 Therefore, a lactose-intolerant individual who consumes 120 mL of cow's milk will ingest 5.6 g of lactose and possibly produce 200 mL of intestinal gas. Infants who consumed pear juice ingested a comparable amount of fructose (7.5 g) along with 2.4 g of sorbitol. This may translate into more intestinal gas than the infant is able to expel, thus causing gas accumulation in the intestine. This may lead to minor abdominal pain, which may cause the infant to be restless, thus translating into more physical activity. Since the gut transient time in 6-month-old infants is approximately 2 hours,22 this may explain why most of the change in physical activity and metabolic rate occurred during the last hour of the study.
Sorbitol is a sugar alcohol that is used as a sugar substitute in many foods.23 Approximately 1% of ingested sorbitol is absorbed though the intestinal wall. Once in the hepatic artery, sorbitol is quickly converted to fructose by L-iditol dehydrogenates in hepatic cells and metabolized normally down the glycolytic pathway.24 The rest of the ingested sorbitol is fermented in the lower intestine by colonic bacteria. The resulting gas production from the colonic fermentation may cause minor discomfort, thus leading to increased physical activity. However, without reliable measurements of minor physical activities, such as those obtained with the infant respiratory chamber,12,13 parents may not be aware of their infant's irritability. The discomfort may not be severe enough for the parents to notice but may be the reason for increased restlessness.
A much higher respiratory quotient was found in infants who had incomplete carbohydrate absorption. This indicates increased carbohydrate utilization. The greater amount of physical activity may be causing increased utilization of stored glycogen during the observation period. Furthermore, some fructose is absorbed, especially in the presence of glucose, and immediately metabolized.24 This may also contribute to the increased respiratoryquotient found in the infants with incomplete carbohydrate absorption.
We have demonstrated that most infants who consume fruit juices containing sorbitol, such as pear juice, are more physically active and have increased carbohydrate malabsorption. It is suggested that parents give their infants non–sorbitol-containing juices. In our study, we found that most of the infants tolerated white grape juice well, as shown by a low amount of physical activity and complete carbohydrate absorption.
Accepted for publication February 10, 1999.
This study was supported by the Maimonides Research and Development Foundation, Brooklyn, NY; Miami Children's Hospital Foundation, Miami, Fla; and Welch Foods Inc, Concord, Mass.
We thank the families of the infants who volunteered their time for our studies.
Editor's Note: So why do so many kids drink apple juice? Or is it only my imagination?—Catherine D. DeAngelis, MD
Corresponding author: Russell Rising, MS, PhD, Research Institute, Miami Children's Hospital, 3100 SW 62nd Ave, Miami, FL 33155.