Feeding behavior in neonates whose diet contained medium-chain triacylglycerols: short-term effects on thermoregulation and sleep

¹ From the Unité de Recherches sur l’Environement Toxique Périnatal–Adaptations Physiologiques et Comportementales (EA 2088), Faculté de Médecine, Université de Picardie Jules Verne (FT, VB, KC, J-PL), and the Service de Néonatologie-Pédiatrie II, CHU-Amiens Nord (AL), Amiens, France.

² Supported by the Regional Council of Picardie and the French Ministry of Research.

³ Address reprint requests to F Telliez, ETP-APC (EA 2088), Faculté de Médecine, Université de Picardie Jules Verne, 3 rue des Louvels, 80036 Amiens Cedex, France. E-mail: frederic.telliez{at}sa.u-picardie.fr.

ABSTRACT  
Background: Feeding formulas for premature infants often contain medium-chain triacylglycerols (MCTs). However, previous studies in animals and adults showed that MCTs may decrease food intake.

Objectives: The objectives were to determine in hospitalized premature infants whether food intake is modified by dietary MCT supply and to assess the effects on thermoregulation and sleep, which are involved in the regulation of energy metabolism and in the optimal physiologic development of the neonates.

Design: Food intake, body mass, and nutritional efficiency during 3 consecutive days were compared in 2 groups of neonates according to the fat composition of their feeding formula [MCT group: 37% MCT, 63% long-chain triacylglycerols (LCTs); LCT group: 100% LCT]. On the third day, sleep and metabolic rate were recorded in the morning during an interval between meals.

Results: Regardless of day, energy intake was greater in the MCT group than in the LCT group ( Conclusion: The ratio of MCTs to LCTs in neonates’ feeding formulas can modify physiologic functions involved in energy-balance regulation.

Key Words: Feeding behavior • medium-chain triacylglycerols • neonate • thermoregulation • sleep

INTRODUCTION  
Milk formulas for premature infants (whose digestive function is immature) often contain medium-chain triacylglycerols (MCTs), ie, those containing saturated fatty acids with carbon chain lengths of 6–10 atoms, which confer nutritional characteristics that differ from those of long-chain triacylglycerols (LCTs) (1). Compared with LCTs, MCTs are more easily hydrolyzed, better absorbed (2), and more rapidly oxidized (3). As a consequence, MCTs are often referred to as a source of readily available energy for premature infants.

However, MCTs could well have adverse effects, especially on feeding behavior. Dietary MCTs decreased short-term food intake in chickens (4), rats (5), and adult humans (6, 7). This effect could be attributed to the fact that MCTs increase the plasma concentrations of cholecystokinin (CCK) (8), a satiety hormone (9). Thus, in neonates, the MCT content of the diet may modify daily energy intake with potential long-term adverse effects on growth. This hypothesis remains the subject of debate.

The type of nutrient could also interfere with thermoregulatory and sleep processes, both of which are implicated in energy-balance regulation. The thermic effect of diet is an important component of body temperature maintenance in premature infants (10, 11). The duration and structure of sleep are also altered by the quality and quantity of food ingested (12–16). Modifications in body temperature and sleep patterns could be of paramount importance in the maturation of neonates’ central nervous system (17).

The present study aimed to determine whether neonates’ food intake is modified by the dietary supply of MCT. Repercussions of the diet on thermic effect and sleep were also assessed.

SUBJECTS AND METHODS  
Subjects and feeding
Seventeen healthy premature neonates, nursed in closed incubators in a hospital pediatric department, were enrolled in this study after their parents were informed of the protocol and had given written, informed consent. The protocol was approved by the Regional Ethics Committee of Picardie. Neonates with neurologic or cardiorespiratory problems or infections, or any combination of those conditions, were not included in the experiment.

Neonates were randomly assigned to 1 of 2 groups (Table 1). No significant differences were observed between the 2 groups with regard to general clinical variables. The neonates were fed with 2 commercially available formulas, currently used in the hospital’s pediatric department, spanning the highest and the lowest rates of MCT supply. The 2 formulas were isoenergetic and isonitrogenous [293 kJ/100 mL energy content, 57.7% carbohydrate (by wt), 14.6% protein, 24.8% fat, and 2.9% mineral components] and differed only in their relative fat composition, ie, the ratio of MCT to LCT (Table 1). The MCT group was fed with a formula enriched with MCTs, containing 37% MCT and 63% LCT (% by wt of total fat). The fat supply for the LCT group was not enriched with MCTs. At the time of the study, no commercial formula used in the pediatric department contained >40% MCT; a 40% component of MCT seems to be the maximum at which an overload of the ß-oxidation capacity in premature infants, and therefore the stimulation of the -oxidative pathway secretion of dicarboxylic acids, can be avoided. The current tendency is to reduce to 0% the amount of MCT in the feeding formulas for premature infants.

View this table:
TABLE 1 . Characteristics of subjects and clinical variables for the 2 study groups and nutritional characteristics of the study formulas¹  
Each neonate was fed with the formula for the 5 d immediately before each experimental session. Feeding nursing care corresponded to the usual procedure and was not modified by the experiment. Each neonate was fed by bottle, with feeding episodes occurring every 3–4 h according to the neonate’s feeding demands, eg, awakening, body activity, and crying. Food was given until the neonate appeared to be satiated, ie, stopped sucking the bottle.

Experimental design
Each infant was studied over 3 consecutive days (day 1–day 3), during which time food intake, body weight, and nutritional efficiency (ie, the ratio of energy supply and body weight gain, calculated over a 24-h period measured from 0800 on day_x to 0800 on day_x+1) were recorded. Each neonate was weighed with accurate scales (±5 g; Sartorius, Palaiseau, France) each morning before feeding. Food intake was estimated from the quantity of milk ingested by the neonate at each feeding episode.

On the third day, sleep and metabolic rate (oxygen consumption, O₂) were recorded in the morning during an interval between meals, ie, from 1 h after a feeding and continuing until spontaneous awakening occurred. This procedure standardized the measurements with regard to the postprandial metabolic state. Gudinchet et al ( Throughout the study, the neonates were nursed at thermoneutrality (19) in a closed, convectively heated incubator (MP4; Médipréma, Chambray-les-Tours, France). Neonates wore only a small diaper and were lying in a supine position on a mattress. The air temperature of the nursery room was controlled and kept constant at 25°C with a relative air humidity of 40%. During the experiments, the neonates were kept under optimal conditions with minimal opening of the incubator portholes and no blood sampling or diaper changing to ensure that their energy requirement was not increased by activity.

Measurements
Metabolic rate ( Esophageal temperature (T_es) was measured with a thermistor probe [YSI 402 (accuracy ± 0.1°C); Bioblock, Illkirch, France] introduced via the mouth and inserted to a distance of 10 cm from the lips. Skin temperature (T_sk) was averaged from 2 local skin temperatures measured with thermocouples [K (accuracy ± 0.15°C); Bioblock], one applied to the right cheek and the other to the righthand side of the abdominal region. Sensors were fixed to the skin surface with rubber tape and covered with aluminum to eliminate the effect of heat radiation reflected from the incubator walls.

The metabolic rate and body temperatures were measured at 10s intervals throughout the experiment. Because of a technical problem, the abdominal skin temperature was not recorded for one neonate.

Sleep stages (active sleep, intermediate sleep, and quiet sleep) were scored for each 30-s period (20) on the basis of electrophysiologic recordings: electroencephalograms from the right and the left Rolando-occipital leads, electrocardiogram, and respiratory signals. Eye movements were monitored with the use of a mechanogram attached to an eyelid (21). Throughout the experiment, visual observation of the neonate’s behavior was made by the same experimenter.

Statistics
Two-factor analysis of variance for repeated measures using the correction of Geisser and Greenhouse (22) and t tests were carried out to test the effects of various amounts of MCT on food intake, nutritional efficiency, metabolic rate, body temperatures, and sleep. Values expressed as percentages have undergone arc sine transformation to stabilize the variance (23). P < 0.05 was accepted as the level of significance. Indicative results (ie, nonsignificant, 0.05 < P < 0.10) are sometimes given when relevant.

RESULTS  
Food intake and nutritional efficiency
No significant differences in digestive physiologic function (regurgitation, presence of residue in the stomach, diarrhea, or constipation) were observed between the groups. The daily energy intake as a function of the 3 consecutive days is shown in Figure 1. No significant interaction between time and type of feeding was observed. It is striking that, throughout the experiment, food intake was significantly higher in the MCT group than in the LCT group (by 67.3 kJ·kg^-1·d^-1; P = 0.007). This accounted for 14–16% of the total energy supply. As shown in Figure 2, the difference in favor of the MCT group resulted from the ingestion of more milk (11.3 kJ·kg^-1·d^-1; P = 0.01) at each feeding episode; the number of feeding episodes each day did not differ significantly (MCT group: 7.1 ± 0.3 episodes; LCT group: 7.4 ± 0.5 episodes).

View larger version (13K):
FIGURE 1. . Mean (±SD) daily energy intake of the neonates in the medium-chain triacylglycerol (MCT, •; n = 10) and long-chain triacylglycerol (LCT,  

View larger version (12K):
FIGURE 2. . Mean (±SD) number of feeding episodes per 24 h and food intake ingested at each feeding episode for the neonates in the medium-chain triacylglycerol (MCT,  
Over the duration of the experiment, intergroup differences in body mass gain (MCT group: 16 ± 13 g/kg; LCT group: 15 ± 11 g/kg) and nutritional efficiency (MCT group: 0.033 ± 0.026 g/kJ; LCT group: 0.036 ± 0.027 g/kJ) were not significant.

The thermic effect of diet
On the third day of the experiment, oxygen consumption was significantly higher (P < 0.001; Figure 3) in the MCT group (7.0 ± 0.5 mL·min^-1·kg^-1) than in the LCT group (5.2 ± 1.1 mL·min^-1·kg^-1), whereas the incubator air temperature did not differ significantly between the 2 groups (MCT group: 33.29 ± 0.22°C; LCT group: 33.25 ± 0.58°C) and remained in the thermoneutral range defined by Sauer et al (19). The level of carbon dioxide production (
View larger version (8K):
FIGURE 3. . Mean (±SD) metabolic rate expressed as oxygen consumption (  
Esophageal temperature tended to be higher in the MCT group (37.12 ± 0.26°C) than in the LCT group (37.03 ± 0.24°C), but the difference was not significant. The intergroup difference was indicative for overall mean skin temperature (MCT group: 36.88 ± 0.24°C; LCT group: 36.66 ± 0.25°C; P = 0.08), but this was due only to a significantly higher cheek skin temperature (0.31°C, P = 0.04) in the MCT group.

Sleep
Total sleep time was significantly (P = 0.01) shorter in the LCT group (-52 min) than in the MCT group (Table 2) as a consequence of an earlier awakening. The time of the sleep onset, wakefulness after sleep onset, and the frequency of sleep stage changes were not modified by the regimen. No significant differences in sleep structure were observed.

View this table:
TABLE 2 . Sleep values for the 2 groups of neonates¹  

DISCUSSION  
Food intake
In the present study, in contrast with the findings reported for adult humans and animals, MCT increased the daily energy intake in comparison with LCT by increasing the average food intake per feeding episode. Various hypotheses, all relating to signals that alter food intake, such as chemical signals from the intestines, mechanical signals from the stomach, and sensory stimulation, can explain this observation.

MCT ingestion elicits the endogenous secretion of CCK in rats (8). Furuse et al (5) used a CCK-A receptor antagonist to show that the inhibitory effect of MCT on food intake in rat is related in part to the CCK satiety signal. In neonates, CCK is produced in response to breast-feeding (24, 25). Marchini and Linden (26) pointed out that CCK concentration is positively correlated with the volume of milk ingested, which suggests that CCK is a major satiety signal and thus is a factor in the regulation of food intake in breast-feeding newborns. On the contrary, Tornage et al (25) did not find any increase in plasma CCK concentrations in bottle-fed, premature infants 3–4 d after birth. They (27) also stressed that plasma CCK did increase after feeding via nasogastric tube but only in combination with kangaroo care (ie, skin-to-skin contact between the mother and the baby). Thus, these findings suggest that physical contact with the mother is a prerequisite for the stimulation of CCK secretion during food intake. This was never the case in the present study.

Feeding nursing care (eg, type of bottle, milk color, type of teat, noise level in the environment) was not significantly different for the 2 groups. This rules out possible bias due to visual, mechanical, or auditory stimulation, which could affect cortical control of feeding behavior. Finally, attention should be paid to preabsorption satiating factors such as smell and taste. When a difference of olfactory or taste sensory stimulation was observed, it favored a preference for LCT consumption in rats (28) and chickens (4, 29). Differences between the viscosity of MCT (25–30 cP) and of LCT (65–75 cP) induce a lower palatability in the MCT regimen than in the LCT regimen (30, 31). Nevertheless, adult humans seem unable to distinguish MCT from corn oil (which predominantly contains LCT) in mixed flavored drinks (7).

As reported above, the effect of MCT on food intake does not appear to be elicited by a CCK-based chemical signal. This observation cannot explain the greater energy intake by the MCT group. Another possible explanation relates to the sensory vagal nervous signals from stomach distension, which project via the solitary tract nucleus in the satiety center within the hypothalamus’s ventromedian nucleus. The type (liquid or solid), energy density, osmolality, and quality (type of carbohydrates and proteins) of the food were the same for each of the 2 groups, and those factors cannot be implicated in a modification of gastric emptying. However, other studies have shown that gastric emptying is more rapid with diets containing a high proportion [50% (32) and 94% (33)] of MCT as fat than with diets containing a low proportion of MCT [14% (32) and 6% (33)], but this difference probably resulted from better direct absorption in the stomach (34). Therefore, the slower emptying occurring with LCT could enhance the filling and the distension of the stomach, inducing satiety more rapidly, as suggested by the lower food intake observed in the LCT group than in the MCT group. Under our experimental conditions, the most plausible explanation of the present results relates to distension-related afferent sensory inputs into the satiety center.

The thermic effect of diet
As a result of the difference between the energy supply (67.3 kJ·kg^-1·d^-1) and the quality of lipids ingested in the 2 groups, energy expenditure (as reflected by the metabolic rate) was greater in the MCT group. This extra energy expenditure induced an increase in cheek skin temperature, but there was no significant change in esophageal and abdominal skin temperatures. The increase in cheek skin temperature—attributable to increased cutaneous perfusion at the face level—could be interpreted as a thermoregulatory process involving selective brain-cooling activation. This mechanism could be particularly important in neonates, because the head is a major site of heat production (40% of energy metabolism) and of heat transfer to the environment (25% of the total body skin surface area) (35, 36). These observations suggest that the effect of MCT on energy expenditure could be related to the activation of the brain metabolism. This is supported by the fact that ketone bodies, the plasma concentration of which is positively correlated with the MCT supply (37), can be used as a major source of energy for developing the neonatal brain (38–40).

Several explanations can be posited for the high value of the respiratory quotient. First, the error of measurement by indirect calorimetry (1%) is sufficient to obtain a respiratory quotient just >1. Second, indirect calorimetry does not take into account the biosynthetic process. When the rate of lipogenesis is higher than that of lipid oxidation (as is the case in a growing neonate), the respiratory quotient is often 1.0–1.1.

Sleep
The reduction in total sleep time in the LCT group was mainly attributed to earlier awakening, and it could be considered an integral part of the behavioral component of a feeding episode. As suggested by Himms-Hagen (41), in a hypothesis of thermoregulatory feeding in newborns, the interval between meals (ie, the duration of sleep before the next meal) depends in part on the amount of the preceding food intake and therefore on the thermic effect of the food. As a result, the lower food intake in the LCT group could be lead to a shortened intermeal interval, which would reduce sleep time. It would be interesting to test this assumption over a 24-h period to ascertain the consequences of feeding behavior on the sleeping-waking cycle in neonates.

Summary
In conclusion, the ratio of MCT to LCT in neonates’ feeding formulas can modify physiologic functions involved in energy-balance regulation. Food intake, energy expenditure, and sleep time were greater in an MCT-fed group than in an LCT-fed group. It is interesting to point out the physiologic consequences of 2 formulas currently used in neonates, which differ only in the type and characteristics of the lipids they supply. This finding could improve the physician’s criteria for choosing the formula according to the neonate’s requirements. For example, when the physician is confronted with an insatiable neonate, it is better to prescribe the LCT formula. In contrast, the MCT formula is better for premature infants who have difficulty in maintaining their body temperature.

ACKNOWLEDGMENTS  
We dedicate this study to Professor Bernard Risbourg, Head of the Department of Pediatrics II of the University Hospital of Amiens, who died recently. We thank Marie-Christine Godefroy, supervisor of the nursing staff of Pediatrics II (CHU Amiens), and wish to acknowledge the work of the nursing staff.

REFERENCES  

1. Bach A, Babayan VK. Medium chain triglycerides: an update. Am J Clin Nutr 1982;36:950–62.
2. Roy CC, Ste-Marie M, Chartrand L, Weber A, Bard H, Doray B. Correction of the malabsorption of the preterm infant with medium-chain triglyceride formula. Pediatrics 1975;86:446–50.
3. Sulkers EJ, Lafeber HN, Sauer PJJ. Quantification of oxidation of medium-chain triglycerides in preterm infants. Pediatr Res 1989;26:294–7.
4. Furuse M, Mabayo RT, Choi YH, Okumura J. Feeding behaviour in chickens given diets containing medium chain triglyceride. Br Poultry Sci 1993;34:211–7.
5. Furuse M, Choi YH, Mabayo RT, Okumura JL. Feeding behavior in rats fed diets containing medium chain triglyceride. Physiol Behav 1992;52:815–7.
6. Van Wymelbecke V, Himaya A, Louis-Sylvestre J, Fantino M. Influence of medium-chain and long-chain triacylglycerols on the control of food intake in men. Am J Clin Nutr 1998;68:226–34.
7. Rolls BJ, Gnizak N, Summerfelt A, Laster LJ. Food intake in dieters and nondieters after a liquid meal containing medium-chain triglycerides. Am J Clin Nutr 1988;48:66–71.
8. Douglas BR, Jansen JB. Effect of various triglycerides on plasma cholecystokinin levels in rats. J Nutr 1990;120:686–90.
9. Gibbs J, Young RC, Smith GP. Cholecystokinin decreases food intake in rats. J Comp Physiol Psychol 1973;84:488–95.
10. Mestyan J, Jarai I, Kekete M, Soltész GY. Specific dynamic action in premature infants kept at and below the neutral temperature. Pediatr Res 1969;3:41–50.
11. Mestyan J, Jarai I, Fekete M. The total energy expenditure and its components in premature infants maintained under different nursing and environmental conditions. Pediatr Res 1968;2:161–71.
12. Fagioli I, Baroncini P, Ricour C, Salzarulo P. Decrease of slow-wave sleep in children with prolonged absence of essential lipids intake. Sleep 1989;12:495–9.
13. Peirano P, Fagioli I, Bhanu Singh B, Salzarulo P. Effect of early human malnutrition on waking and sleep organization. Early Hum Dev 1989;20:67–76.
14. Yogman MW, Zeisel S. Diet and sleep patterns in newborn infants. N Engl J Med 1983;309:1147–9.
15. Salzarulo P. Sleep, nutrition and metabolism during development. In: Pollitt E, Amante P, eds. Energy intake and activity. New York: Alan R Liss, 1984:323–7.
16. Fagioli I, Ricour C, Salomon F, Salzarulo P. Weight changes and sleep organisation in infants. Early Hum Dev 1981;5:395–9.
17. Tolaas J. REM sleep and the concept of vigilance. Biol Psychiatry 1978;13:135–48.
18. Gudinchet F, Schutz Y, Micheli JL, Stettler E, Jéquier E. Metabolic cost of growth in very low birth weight infants. Pediatr Res 1982;16:1025–30.
19. Sauer PJJ, Dane HJ, Visser HKA. New standards for neutral thermal environment of healthy very low birthweight infants in week one of life. Arch Dis Child 1984;59:18–22.
20. Curzi-Dascalova L, Mirmiran M. Manual of methods for recording and analyzing sleep-wakefulness states in preterm and full-term infant. Paris: les Editions INSERM, 1996.
21. Pajot N, Vicente G, Dreyfus-Brisac C. Techniques d’enregistrement des mouvements oculaires chez le nouveau-né: comparaison des méthodes. (Techniques of recording eye movement in neonates: a comparison of methods.) J Electrophysiol Technol 1976;2:29–38 (in French).
22. Geisser S, Greenhouse SW. An extension of Bor’s results on the use of the F distribution in multivariate analysis. Ann Math Stat 1958;29:885–91.
23. Winer G. Statistical principles in experiments design. New York: McGraw Hill, 1971.
24. Uvnas-Moberg K, Marchini G, Winberg J. Plasma cholecystokinin concentrations after breast feeding in healthy 4 day old infants. Arch Dis Child 1993;68:46–8.
25. Tornhage CJ, Serenius F, Uvnas-Moberg K, Lindberg R. Plasma somatostatin and cholecystokinin levels in response to feeding in preterm infants. J Pediatr Gastroenterol Nutr 1998;27:199–205.
26. Marchini G, Linden A. Cholecystokinin, a satiety signal in newborn infants? J Dev Physiol 1992;17:215–9.
27. Tornhage CJ, Serenius F, Uvnas-Moberg K, Lindberg T. Plasma somatostatin and cholecystokinin levels in preterm infants during kangaroo care with and without nasogastric tube-feeding. J Pediatr Endocrinol Metab 1998;11:645–51.
28. Edens NK, Friedman MI. Response of normal and diabetic rats to increasing dietery medium-chain triglyceride content. J Nutr 1984;114:565–73.
29. Mabayo RT, Okumura J, Hirao A, Sugita S, Sugahara K, Furuse M. The role of olfaction in oil preference in chicken. Physiol Behav 1996;59:1185–8.
30. Greenberg NJ, Skillman TG. Medium-chain triglycerides. N Engl J Med 1969;320:1045–58.
31. Ruppin DC, Middleton WR. Clinical use of medium chain triglycerides. Drugs 1980;20:216–24.
32. Sidebottom R, Curran JS, Williams PR, Kanarek KS, Bramson RT. Effects of long-chain vs medium-chain triglycerides on gastric emptying time in premature infants. J Pediatr 1983;102:448–50.
33. Siegel M, Krantz B, Lebenthal E. Effect of fat and carbohydrate composition on the gastric emptying of isocaloric feedings in premature infants. Gastroenterology 1985;89:785–90.
34. Hamosh M, Scanlon JW, Ganot D, Likel M, Scanlon KB, Hamosh P. Fat digestion in the newborn. J Clin Invest 1981;67:838–46.
35. Stothers JK. Head insulation and heat loss in the newborn. Arch Dis Child 1981;56:530–4.
36. Marks KH, Devenyi AG, Bello ME. Thermal head wrap for infants. J Pediatr 1985;107:956–9.
37. Wu PYK, Edmond J, Auestad N, Rambathla S, Benson J, Picone T. Medium-chain triglycerides in infant formulas and their relation to plasma ketone body concentrations. Pediatr Res 1986;20:338–41.
38. Gonzales LW, Geel SE. Enhanced cerebral protein synthesis in developing hypothyroid rats: evidence of a delayed maturation. J Neurochem 1978;31:1239–46.
39. Settergren G, Lindblad BS, Persson B. Cerebral blood flow and exchange of oxygen, glucose, ketone bodies, lactate, pyruvate, and amino acids in infants. Acta Paediatr Scand 1976;65:345–53.
40. Kraus H, Schlenker S, Schwedesky D. Developmental changes of cerebral ketone body utilization in human infants. Hoppe Seylers Z Physiol Chem 1974;355:164–70.
41. Himms-Hagen J. Does thermoregulatory feeding occur in newborn infants? A novel view of the role of brown adipose tissue thermogenesis in control of food intake. Obes Res 1995;3:361–9.