Acta Pñ diatr 90: 196± 8.

2001

Hydrolysed protein accelerates the gastrointestinal transport of formula

in preterm infants
WA Mihatsch1 , J Högel2 and F Pohlandt1
Division of Neonatolog y and Pediatric Critical Care1 , Department of Pediatrics and Department for Biometry and Medical
Documentation 2 , University of Ulm, Ulm, Germany

Mihatsch WA, Högel J, Pohlandt F. Hydrolysed protein accelerates the gastrointestinal transport of
formula in preterm infants. Acta Pædiatr 2001; 90: 196–8. Stockholm. ISSN 0803-5253
Vomiting, large gastric residuals and abdominal distension are common in very immature infants
on formula feeding. The present trial investigated whether a protein hydrolysate formula reduces
the gastrointestinal transit time in preterm infants. Fifteen preterm infants (median gestational age
29 (24–32) wk, birthweight 1241 (660–1900) g, postnatal age 18 (5–54) d) on full enteral feeds
(> 150 ml/kg*d) were enrolled. It was hypothesized that the gastrointestinal transit time is at least
2 h shorter when protein hydrolysate formula is fed compared with standard preterm formula. In a
randomized cross-over design study, each formula was fed for 5 d. On days 4 and 9 the
gastrointestinal transit time was estimated using carmine red. The protein hydrolysate formula had
a markedly shorter gastrointestinal transit time (9.8 h) than the standard formula (19 h)
(p = 0.0022, two-sided Mann-Whitney U test).
Conclusion: The hydrolysate protein formula accelerated gastrointestinal transit of milk and stools,
but whether hydrolysate formulas enable a more rapid establishment of full enteral feeding in
preterm infants needs to be investigated.
Key words: Gastrointestinal transit time, infant nutrition, preterm infant, protein hydrolysate
formula
Frank Pohlandt, Universitäts-Kinderklinik, 89070 Ulm, Germany (Tel. ‡49 731 5002 7740, fax.
‡49 731 5002 6723, e-mail. frank.pohlandt@medizin.uni-ulm.de )

In full-term neonates and infants, gastric emptying is
faster when a protein hydrolysate formula is fed as
compared with formulas containing native cow’s milk
protein (1, 2). Whether hydrolysate formulas might
have a shorter gastrointestinal transit time (GTT) is
unclear (3).
The present study investigates whether a protein
hydrolysate formula (Formula H) has a shorter GTT in
preterm infants than a standard preterm formula
(Formula S).
milk protein with a whey to casein ratio of 60:40 (Table
1). Formula H contained an ultra-Ž ltrated mixture of
hydrolysed whey and hydrolysed casein, with a whey to
casein ratio of 60:40 (molecular weight: 75% < 1500
Dalton, 15% free amino acids) and contained more
protein (23 g/l) and minerals because of their reduced
bioavailability in hydrolysate formulas (4). There were
only minor differences between the amino acid compo-
sitions of the proteins (5). Both formulas were supplied
in ready-to-feed form (Milupa GmbH, 61381 Frie-
drichsdorf, Germany).

Patients and methods

Study population
Fifteen preterm infants on full enteral feeding
(> 150 ml/kg*d) were enrolled. Mother’s milk intake
(one infant) up to 20% of the total daily intake was
accepted. Written informed parental consent was
obtained and the study protocol approved by the
institutional ethics committee. Anomalies which might
interfere with nourishing were exclusion criteria.
Formulas
The standard preterm formula contained native cow’s
Study design
The infants were fed according to their individual
ability (bottle feeding, n = 7, gavage feeding, n = 8). In a
cross-over design, the Ž rst formula, allocated at
random, was fed from days 1 to 5 and the alternate
formula from days 6 to 10 of the study. On days 4 and 9,
the 0 a.m. feed was stained with 10 mg carmine red
(Micro-Plus, 37627 Stadtoldendorf, Germany). The
time from feeding the dye to its appearance in the
diaper was recorded as GTT. The fecal marker was
identiŽ ed by the attending nurse, who checked the
diapers at least every 2 h (inter-observer variability has

Ó 2001 Taylor & Francis. ISSN 0803-525 3

ACTA PÆDIATR 90 (2001)
Table 1. Composition of the study formulas.
Hydrolysate formula (Formula H)
Energy density (kJ/l) 3350
Osmolarity (mOsmol/l) 260
Protein (g/l) 23 (hydrolysed)
Lipid (g/l) 40
Medium-chain triglycerides 12.1%
Carbohydrates (g/l) 82
Lactose (g/l) 53
Maltodextrin (g/l) 29
Minerals (g/l) 5 4
Phosphate (mg/l) 510
Calcium (mg/l) 940
not been studied). Feeding volumes, gastric residuals (in
case of gavage feeding), regurgitation or vomiting, and
number and quality of stools (hard, soft or liquid) were
recorded as secondary outcome variables. Weight gain
was monitored. During the 10-d study period, supple-
mentation with calcium or phosphate was stopped.
There were no enemas and no laxatives during the
study.
Data analysis
It was estimated that a difference in transit time
(primary outcome) between the two formulas of less
than 2 h would be of no importance clinically. The study
hypothesis H1 was deŽ ned as follows: The GTT of the
hydrolysate formula is at least 2 h shorter than the GTT
of the standard preterm formula.
There were no data available on the transit times and
the standard deviations of the two formulas. Therefore,
an adaptive interim analysis (internal pilot study) (6) of
the primary outcome was scheduled to be made as soon
as the results of 15 infants were available (two-sided
Mann-Whitney U test). The adaptive interim analysis
was chosen to estimate the variance of the transit time
and then to calculate the Ž nal sample size or to Ž nish the
trial in case of extremely pronounced effects. In the
interim analysis of the study presented, H0 would be
rejected if the difference between the formulas reached
a* = 0.0087. This level of signiŽ cance of the adaptive
interim analysis guarantees an overall a level of 5% for
the Ž nal combination test.
An explorative investigation of carry-over and period
effects was carried out (7). Wilcoxon tests and the chi-
squared test were used for secondary outcome analysis
Table 2. Gastro-intestina l transit time in hours measured by carmine red passage.
All infants
N 15
Transit time Formula H 9.8 (5–20.8)
Transit time Formula S 19 (6–66)
Difference in transit time (Formula S – Formula H) 12 (¡3.75–45.3)
Data are shown as median (minimum – maximum). Formula H, hydrolysed protein formula; Formula S, standard preterm formula (native protein).
Hydrolysed protein accelerates transpor t 197
Standard preterm formula (Formula S)
2980
215–220
20 (non-hydrolysed )
35
7.5%
77
50
27
350
700
(level of signiŽ cance p < 0.05, Statistical Systems for
Personal Computers Package, SAS Institute, Cary, NC).
Data are presented as median, minimum and maximum.
Results
Fifteen preterm infants (median gestational age 29 (24–
32) wk, postmenstrual age at study entry 32 (29–35) wk,
body weight at study entry 1440 (1020–2070) g) were
enrolled. No participants withdrew and there were no
adverse events.
The daily feeding volumes were the same with both
formulas: Formula H, 163 ml/kg*d (range 150–222
ml/kg*d); Formula S, 165 ml/kg*d (range 141–208
ml/kg*d). At the interim analysis (Table 2), the median
individual difference in GTT was 12 h (p = 0.0022). H0
was rejected and the study stopped. No carry-over or
period effects were observed. In addition, in the Ž rst
period (parallel comparison) Formula H had a shorter
GTT than Formula S (p < 0.05, Mann-Whitney U test).
Weight gain was equal with both formulas: Formula H,
22 g/d (range 2–52 g/d); Formula S 22 g/d (range 12–46
g/d).
There was no difference between the two formulas
with regard to the frequency of vomiting/spitting and
the average volume of gastric residuals (data not
shown). On study days 4 and 9 there was no signiŽ cant
difference between the two formulas with regard to the
number of stools per infant (5 (1–8) vs 4 (0–7); Formula
H vs Formula S) and the total number of stools (67 vs
54), but soft stools were reported more often with
Formula H feeding (5 hard stools, 62 soft stools) than
with Formula S feeding (21 hard stools, 33 soft stools).
First formula was Formula H First formula was Formula S
8 7
9.4 (5–18) 12.2 (6–20.8)
21.8 (6–44) 15.5 (8.3–66)
13.2 (¡3.75–33.8) 8.8 (1.3–45.3)
198 WA Mihatsch et al.
Discussion
This is the Ž rst time that the GTT of two different
formulas has been investigated in preterm infants in a
randomized cross-over design as primary outcome. The
study design justiŽ es the conclusion that the reported
difference in transit time is mainly caused by the
different proteins, but several minor differences be-
tween the formulas need to be discussed.
The strength of Formula H was enriched by roughly
10%. Increasing caloric density delays gastric emptying
(8), but there are no studies on gastrointestinal transit
time available. Nearly all existing data on transit time
measurements in preterm infants are part of metabolic
studies where carmine red staining was used for the
identiŽ cation of stool collection periods. In preterm
infant formulas long-chain triglycerides tend to slow
down gastric emptying (9). However, medium-chain
triglyceride portions above 40% of the fat content are
associated with longer transit times (10, 11). If at all, the
higher medium-chain triglyceride content of Formula H
should have prolonged GTT. Certain fatty acids (C14:0,
C16:0 and C18:1) are positively correlated with stool
hardness (12). In Formula H their overall concentration
and the total lipid concentration were higher, but stools
were not harder and transit time was faster. Hydrolysis
of the protein increased the osmolarity and osmotic
effects might have reduced GTT. The shorter GTT of
Formula H could be a re ection of an impaired nutrient
absorption rate, too (4), because there was no difference
between the two formulas with regard to plasma amino
acid concentrations and growth (5), despite the higher
protein and energy intake. However, the median urea
concentration was signiŽ cantly higher with Formula H
feeding, possibly indicating a higher absorption (5).
Finally, a casein-predominant formula has a longer
GTT than a whey-predominant formula in infants (13)
and rats (14). In rats, this difference could be abolished
by pretreatment with the opiate-receptor antagonist
naloxone (14). Mammalian milk contains highly spe-
ciŽ c and potent m-type opioid receptors agonists (e.g. b-
casomorphins) (15). It is a well-known side effect of
opioids that they reduce bowel activity and cause
constipation. In the present study, the whey to casein
ratio of both formulas was equal, but hydrolysis of the
protein might have reduced the activity of milk protein-
derived opioid receptor agonists and decreased transit
time.
Hydrolyzed protein formula feeding was associated
with more soft stools and a tendency towards a higher
ACTA PÆDIATR 90 (2001)
number of stools; however, the power to detect
differences was small because of the small number of
infants.
In conclusion, the present study revealed a reduced
transit time for the hydrolysate formula. We speculate
that enteral feedings can be more rapidly advanced in
preterm infants using the hydrolysate formula.
References
1. Tolia V, Lin C-H, Kuhns LR. Gastric emptying using three
different formulas in infants with gastroesophagea l re ux. J
Pediatr Gastroentero l Nutr 1992; 15: 297–301
2. Billeaud C, Guillet J, Sandler B. Gastric emptying in infants with
or without gastro-oesophagea l re ux accordin g to the type of
milk. Eur J Clin Nutr 1990; 44: 577–83
3. Sievers E, Santer R, Oldigs H-D, Haase S, Schleyerbac h U,
Schaub J. Gastrointestinal e Passagezeiten bei Frühgeborenen .
Monatssch r Kinderheilk d 1995; 143: S76–80
4. Rigo J, Salle BL, Picaud J-C, Putet G, Senterre J. Nutritional
evaluation of protein hydrolysat e formulas. Eur J Clin Nutr 1995;
49 Suppl 1: S26–39
5. Mihatsch WA, Pohlandt F. Protein hydrolysat e formula main-
tains homeostasi s of plasma amino acids in preterm infants. J
Pediatr Gastroentero l Nutr 1999; 29: 406–10
6. Bauer P, Köhne K. Evaluation of experiment s with adaptive
interim analyses . Biometrics 1994; 50: 1029–41
7. Lehmacher W. Analysis of the cross-over design in the presence
of residual effects. Stat Med 1991; 10: 891–9
8. Siegel M, Lebenthal E, Kranz B. Effect of caloric density on
gastric emptying in premature infants. J Pediatr 1984; 104: 118–
22
9. Siegel M, Krantz N, Lebenthal E. Effect of fat and carbohydrat e
compositio n on the gastric emptying of isocalori c feedings in
premature infants. Gastroenterolog y 1985; 89: 785–800
10. Okamoto E, Muttart CR, Zucker CL, Heird WC. Use of medium-
chain triglyceride s in feeding the low birthweight infant. Am J
Dis Child 1982; 136: 428–31
11. Bergmann RL, Jäger H, Kern M, Bergmann KE. Mittelkettig e
Triglycerid e in der Ernährung von Frühgeborenen . Monatssch r
Kinderheilkd 1977; 125: 953–8
12. Quinlan PT, Lockton S, Irwin J, Lucas AL. The relationshi p
between stool hardness and stool composition in breast- and
bottle-fed infants. J Pediatr Gastroenterol Nutr 1985; 20: 81–90
13. Brown GA, Berger HM, Brueton MJ, Scott PH, Wharton BA.
Nonlipid formula component s and fat absorption in the low
birthweigh t newborn. Am J Clin Nutr 1989; 49: 55–61
14. Daniel H, Vohwinkel M, Rehner G. Effect of casein and beta
casomorphins on gastro-intestina l motility in rats. J Nutr
1990;120: 252–7
15. Teschemache r H, Koch G, Brantl V. Milk protein derived opioid
receptor ligands. Biopolymers 1997; 43: 99–117
Received April 5, 2000; revision received Sept. 7, 2000; accepted
Sept. 7, 2000