Pediatrics International (2007) 49, 156160

doi: 10.1111/j.1442-200X.2007.02336.x

Original Article

Hindmilk and weight gain in preterm very low-birthweight infants

AMALI-ADEKWU OGECHI,1,2 OGALA WILLIAM1,3 AND BODE-THOMAS FIDELIA1,2
1
Department of Paediatrics, Jos University Teaching Hospital, 2Department of Paediatrics, Faculty of
Medical Sciences, University of Jos, Jos and 3Department of Paediatrics, Faculty of Medicine, Ahmadu
Bello University, Zaria, Nigeria
Abstract

Background: Feeding of own mothers milk to preterm very low-birthweight infants often results in suboptimal
weight gain in these infants for whom energy requirements are high but in whom volume tolerance is limited.
Therefore the purpose of the present paper was to investigate the effect of selective hindmilk feeding on the
growth of preterm very low-birthweight babies.
Methods: Preterm very low-birthweight babies admitted into the Special Care Baby Unit of the Jos University
Teaching Hospital, Nigeria between April 2000 and July 2001 were randomized to hindmilk and composite
breast milk feeding for 2 weeks. End-points were weight, occipitofrontal head circumference and length.
Results: For small-for-gestational-age babies, the infants fed on hindmilk gained weight at a mean rate of 12.92
10.95 g/kg per day as compared with a mean rate of 5.01 17.37 g/kg per day for their controls on composite
milk (P < 0.0001). For appropriate-for-gestational-age babies, the mean rate of weight gain for the hindmilk
group was 12.99 10.75 g/kg per day while that for their controls on composite milk was 8.29 20.56 g/kg per
day (P < 0.01). There were no significant differences in the rates of increase in length and occipitofrontal head
circumference between the groups. The lipid content of the hindmilk was 1.6-fold that of composite milk.
Conclusion: Preterm very low-birthweight babies fed hindmilk had a higher rate of weight gain compared to
those fed composite milk. It is recommended that the hindmilk fraction of expressed breast milk be predominantly used for the feeding of preterm very low-birthweight babies while in hospital to help shorten their duration of hospital stay.

Key words

hindmilk, low birthweight, preterm, weight gain.

Infant mortality remains high in developing countries, with a

reported rate of 100/1000 live births in Nigeria in the year
2003.1 Approximately 50% of these deaths occurred in the neonatal period.1 Low birthweight accounts for 20 and 36% of neonatal and perinatal mortalities, respectively.2 Recent evidence
shows that feeding of preterm infants in the early weeks after
birth has a major effect on their long-term outcome.3,4 The adjustment of the preterm infant to sudden exposure to the extrauterine environment might depend even more strongly on
provision of own mothers milk than that of the full-term infant. Human milk provides the infants not only with nutrients
needed for growth but also with a large number of bioactive
components that modulate neonatal development.4,5
Benefits such as protection from necrotizing enterocolitis
and other infections, greater enteral feeds tolerance, reduced

Correspondence: Dr Fidelia Bode-Thomas, Department of Paediatrics, Faculty of Medical Sciences, University of Jos, Jos, Plateau
State, Nigeria. Email: bodefide@yahoo.com
Received 10 January 2005; revised 3 November 2005; accepted
16 January 2006.

risk of later allergy, improved retinal function and enhanced

neurocognitive development have been reported with the use of
human milk.4,5 However, the feeding of own mothers milk
often results in suboptimal weight gain for the smaller preterm
infant for whom energy requirements are high but in whom volume tolerance limited.6 Fat is a major energy source for the
newborn. Breast milk lipids provide approximately 50% of the
calories in the milk.7,8 The lipid content of breast milk increases
during a single feeding from foremilk to hindmilk.9 The compositions of foremilk and hindmilk have been examined. Foremilk
is watered down milk, which is thirst-quenching, high in lactose
but low in fat and protein. No differences were observed
between the fractions for the contents of nitrogen, sodium, calcium, phosphorous and potassium.10 Small differences were
observed in copper and zinc, each declining by approximately
56% from foremilk to hindmilk, but such small differences
have been found to have no adverse biological effects.10
Recent research has focused on the very long-chain fatty
acids, arachidonic acid and decosahexanoic acid (22:6n-3),
derivatives of linoleic and linolenic acids, respectively, which
are present in hindmilk.11

Hindmilk and weight gain in VLBW infants

Arachidonic and decosahexanoic acids are components of
phospholipids found in brain and red cell membranes: they
have been associated functionally with improved cognition,
growth and vision.12
Foremilk and hindmilk fractions represent 40 and 60%,
respectively, of daily milk volumes. Lactoengineering strategies in which own mothers milk is fragmented into two portions, foremilk and hindmilk, with selective hindmilk feedings,
have been used extensively in developed countries to provide
very low-birthweight (VLBW) infants with higher energy,
thereby providing a higher calorie-rich feed without the danger of volume overload. This difference in fat composition has
not been utilized in the clinical management of low-birthweight infants in Nigeria. Available data on the subject from
Nigeria are limited to an earlier trial at the Special Care Baby
Unit of the Jos University Teaching Hospital (SCBU JUTH),
Nigeria, by Slusher et al.,13 which reported increased rate of
weight gain in low-birthweight infants while being fed hindmilk. That study, however, involved moderately low-birthweight infants and did not involve controls. The present study
was therefore aimed at comparing the effects of selective hindmilk feeding and feeding of the usual composite breast milk
on the growth rate of preterm VLBW infants.

Methods
Subjects

This randomized controlled trial was carried out at SCBU

JUTH over a period of 14 months (April 2000July 2001). The
ethics committee of the hospital approved the study. Parents
gave informed consent. The subjects were healthy consecutive
preterm VLBW babies admitted into the Unit, who met the
following inclusion criteria : (i) birthweight between 1000 and
1499 g; and (ii) gestational age <37 completed weeks
The exclusion criteria included: (i) presence of any congenital abnormalities; (ii) presence of serious postnatal illness
such as severe birth asphyxia, neonatal sepsis, severe neonatal
jaundice and requirement for prolonged respiratory assistance;
(iii) known maternal HIV infection; and (iv) inability to tolerate full enteral feeding by 72 h of age.
The babies were randomly assigned to either hindmilk
feeding or composite milk feeding groups, with the composite
milk group serving as control. Gestational age was assessed
clinically using the Dubowitz et al. method.14 To control for
the influence of the state of intrauterine nutrition on early postnatal growth, the randomization was further stratified by birthweight and gestational age into four subgroups: small for
gestational age (SGA) on composite breast milk; SGA on
hindmilk; appropriate for gestational age (AGA) on composite
breast milk; and AGA on hindmilk. Each baby was admitted
into a conventional incubator and nursed in a thermoneutral

157

environment with an average temperature of 34C and humidity of 6080%.

Breast milk collection

Breast milk was obtained mechanically from the mothers

using a breast pump into clean, clear containers. Foremilk was
defined as the milk collected for the first 3 min after flow
began.15 A color difference from white to yellow was noted at
the beginning of hindmilk collection. The remainder of the
milk collected until complete emptying of the breast was defined as hindmilk. The fat concentration of the foremilk, hindmilk and composite milk was estimated daily by means of a
creamatocrit16 and the corresponding caloric values of the various creamatocrits were read off a standard table designed by
Lucas et al.16 All the containers for milk expression and collection were washed with a solution of 3.5% sodium hypochlorite
and boiled for 10 min before each use.
Feeding methods

All the babies were fed for the first 4 days after the establishment of enteral feeding with composite milk to ensure uniform
colostrum ingestion. The babies were commenced either on
hindmilk or composite milk depending on the randomly assigned group, on the fifth day, at 100 mL/kg per day and fed
2 hourly by intermittent gavage using nasogastric tube feeding set consisting of 1020 mL syringe barrels connected to
five French gauge feeding tubes.
The volume of feeds was increased daily by 15 mL/kg up to
a maximum of 200 mL/kg per day.9 Each baby was fed for 14
days, after which all babies were reverted to full pumping milk
until discharge. Babies who developed complications such as
abdominal distension, vomiting, apneic attacks or who died
during the course of the study were excluded from analysis.
Anthropometry

The babies were weighed naked daily on a battery-operated

digital electronic infant weighing scale (Baby weigh, Medela,
Crystal Lake, IL, USA), which measured up to the nearest
gram.
Occipitofrontal head circumference and recumbent length
were measured weekly for each baby.
Blood samples

One milliliter of venous blood was collected through a peripheral vein from each baby under aseptic conditions weekly and
analyzed for serum bicarbonate. Metabolic acidosis was diagnosed if there was a base deficit >8 mmol/L and corrected with

158

O Amali-Adekwu et al.

8.4% sodium bicarbonate, which was added to the breast

milk.17,18 The doses were calculated based on the formula:19
0.3weight(kg)base deficit
= millimoles of bicarbonate required.

Data analysis

The data obtained were analyzed using SPSS version 8 (SPSS,

Chicago, IL, USA). Continuous variables were expressed as
means SD. Students t-test and one-way analysis of variance
(anova) was used to compare means. P < 0.05 was accepted
as statistically significant. Linear regression analysis was used
to asses the relationship between weight gain and milk
fraction.

Results
During the study period, a total of 751 babies were admitted
into SCBU JUTH. Three hundred and eighty-eight (51.7%)
of these were male and 363 (48.3%) were female, giving an
M : F ratio of 1.1:1.
Preterm VLBW babies accounted for 99 (13.2%) of all
admissions. Forty-two of these were male and fifty-seven were
female (M : F ratio, 1:1.3). Among the preterm VLBW babies
were five sets of twins (10.1%).
Seventy-seven consecutive preterm VLBW babies were
recruited into the study. Sixty-eight of them completed 14

days of feeding. Five babies, two on composite milk and three

on hindmilk, were excluded from the study because they
developed abdominal distension and vomiting. Three babies
on composite milk and one on hindmilk developed apneic
attacks and were also excluded. All the nine babies subsequently died, probably from necrotizing enterocolitis (55.6%)
and apnea (44.4%). Of the 68 babies who completed the study,
37 were male and 31 were female giving an M : F ratio of
1.2:1.
The birthweight of the babies ranged from 1000 to 1499 g
with a mean SD of 1339.4 144 g. Tables 1 and 2 show the
baseline characteristics of the babies studied.

Milk volume and composition

The average daily milk production was significantly higher

(P < 0.0001) in the hindmilk group (356.33 80.25 mL) than
in the composite milk group (218.64 47 mL) in both SGA
and AGA infants. As shown in Table 3, foremilk and hindmilk
fractions made up 35.1% and 64.9% respectively of the 24 h
milk volumes.
The lipid content of hindmilk was also found to be significantly higher than that of composite milk in both the SGA
and AGA infants. Composite milk had a mean daily creamatocrit of 5.73 1.4% while that of hindmilk was 9.23 1.89%
(P < 0.0001). The caloric content of hindmilk was significantly higher than that of composite milk and foremilk as
shown in Table 3 (P < 0.0001).

Table 1 Baseline characteristics of the 34 preterm VLBW SGA infants (mean SD)
Characteristic
Birthweight (g)
Gestational age (weeks)
Age on enrolment (days)
Weight on enrolment (g)
Length on enrolment (cm)
Head circumference (cm)

Babies fed on hindmilk (n = 17)

Babies fed on composite milk (n = 17)

1278.3 168.2
33.05 2.06
6.05 0.80
1173.45 146.57
37.9 1.49
29.0 1.27

1267.8 149.4
33.97 1.15
5.82 0.70
1191.29 137.06
36.3 2.21
28.2 1.12

0.61
5.049
2.796
1.226
2.502
1.766

0.092
0.542
0.779
0.311
0.306
0.807

SGA, small for gestational age; VLBW, very low-birthweight.

Table 2 Baseline characteristics of the 34 preterm VLBW AGA infants (mean SD)
Characteristic
Birthweight (g)
Gestational age (weeks)
Age on enrolment (days)
Weight on enrolment (g)
Length on enrolment (cm)
Head circumference (cm)

Babies fed on hindmilk (n = 17)

Babies fed on composite milk (n = 17)

1380.47 103.52
32.38 1.68
6.22 0.79
1297.40 144.12
36.62 1.84
28.54 0.88

1430.88 54.42
32.81 0.81
5.38 0.62
1329.76 97.01
37.18 1.31
29.75 0.95

5.62
2.93
10.737
2.314
1.021
3.865

0.04
0.421
0.352
0.021
0.392
0.991

AGA, appropriate for gestational age; VLBW, very low-birthweight.

Hindmilk and weight gain in VLBW infants

159

Table 3 Volumes and composition in a 24 h milk collection (mean SD)

Parameter

Volume (mL)
Volume %
Energy (Joules/mL)
Creamatocrit (%)

FM

HM

CM

CM vs HM
P

CM vs HM
P

124.84 39.12
35.1
2.6 0.34
4.7 0.86

231.5 53.9
64.9
3.73 0.50
9.23 1.89

356.3 80.3
100
2.8 0.38
5.73 1.4

<0.0001

<0.0001

<0.000
<0.000

<0.0001
<0.0001

CM, composite milk; FM, foremilk; HM, hindmilk.

Weight gain

The rate of weight gain was significantly higher in the babies

fed on hindmilk than the controls for both SGA and AGA infants, as shown in Figure 1.
The mean daily weight gain for the SGA infants on hindmilk was 12.92 10.95 g/kg per day with a range of 1.2 to
21.6 g/kg per day as compared with a range of 15.2 to 24.2
g/kg per day and a mean rate of 5.01 17.37 g/kg per day for
the infants on composite milk (P < 0.0001).
The AGA babies on hindmilk had a mean daily weight gain
of 12.99 10.75 g/kg per day and a range of 12.2 to 28.4 g/
kg per day (P = 0.009).
The occipitofrontal head circumferences and lengths did
not increase significantly from 1 week to the other in any of
the groups.
There was a positive correlation between the rate of weight
gain and the lipid content (creamatocrit) of the breast milk in
both AGA and SGA infants (R = 0.59).
Significant metabolic acidosis, base deficit > 8 mmol/L
was present in seven (20.6%) of the 34 patients fed on hind18

Mean weight gain (g/kg/d)

16
14
12
10

Week 1
Week 2

8
6
4
2
0
Hindmilk
AGA

Hindmilk
SGA

Composite
milk AGA

Composite
milk SGA

Sample group

Fig. 1 Weekly weight gain by babies feed on hindmilk and

composite milk.

milk and in nine (26%) of the 34 patients fed composite milk.

The lowest serum bicarbonate level (10 mmol/L) occurred in
one of the SGA babies fed with composite milk. There was no
significant difference in the frequency of metabolic acidosis
between the babies fed on composite milk and those fed on
hindmilk (P = 0.36).

Discussion
This study has clearly shown that preterm VLBW babies fed
hindmilk gain weight significantly faster than those fed composite milk. The more rapid rate of weight gain of infants on
hindmilk obtained in the present study is very likely as a result
of the higher lipid and therefore higher caloric content of
hindmilk. This observation compares favorably with results
obtained in an earlier study at SCBU JUTH by Slusher et al.13
and also with the findings of Valentine et al.7 and Schanler.10
The increased rates of weight gain noted in the same babies on
the commencement of hindmilk feeding by these authors
suggested that the increments resulted from the increased fat
absorption and deposition that occurred with the greater fatcontaining hindmilk feeds. This conclusion was further
strengthened by the fact that Valentine et al. reported a slowing
of weight gain when mothers had to revert to composite milk
on completion of their study.7
The findings in the present study, however, differed from
the work by Spencer et al.20 in Nottingham, England, where
they observed no significant increment in the rate of weight
gain in babies fed hindmilk. This observation may be due to
their very small sample size of eight patients, with only two of
them receiving hindmilk.
The insignificant rates of increase in length and occipitofrontal head circumference noted in the present study from
week 1 to week 2, were similar to reported growth rates of
these parameters from other studies on growth and energy
intake in preterm babies.7,15 It is known that the rate of growth
in length of the fetus peaks in the second trimester, unlike that
of weight, which peaks in the third.21
The lipid content of the hindmilk found in the present study,
twice that of foremilk and 1.6-fold that of composite milk, is

160

O Amali-Adekwu et al.

in agreement with values of 23-fold higher hindmilk than

foremilk reported from previous studies.3,9,13,17
The rate of development of metabolic acidosis did not differ significantly between the hindmilk and composite milk
groups in the present study. The lower rates of weight gain in
the composite milk groups could not therefore be attributed to
this known cause of growth faltering.17
Early establishment of adequate caloric supply is very
important in the successful adaptation of preterm VLBW
infants to extrauterine life. The present study has demonstrated
that the technique of feeding only the hindmilk fraction of
expressed breast milk can provide more calories in a relatively
smaller volume of milk. In addition to ensuring optimal weight
gain in these babies, this would decrease the risk of aspiration
of feeds and other problems associated with volume overload
in preterm VLBW infants. The faster rate of weight gain in
these infants would shorten their duration of hospitalization
and the risks of acquisition of nosocomial infections. This
would ultimately result in the reduction of morbidity and mortality in these infants.

References
1 UNICEF. State of the Worlds Children. 2005.
2 Omene JA, Okolo AA, Imodeheme D, Omu A. Trends in perinatal mortality at the University of Benin Teaching Hospital
Nigeria. East Afr. Med. J. 1984; 61: 4619.
3 Lucas A, Morley RM, Cole TJ, Lister G, Leeson-Payne C.
Breastmilk and subsequent intelligence quotient in children
born preterm. Lancet 1992; 339: 2614.
4 Lucas A, Morley RM, Cole TJ. Early diet in preterm babies and
developmental status at 18 months. Lancet 1990; 335: 14778.
5 Lucas A, Cole TJ. Breastmilk and necrotizing enterocolitis.
Lancet 1990; 336: 151923.
6 Atkinson SA, Radde IC, Anderson GH. Macromineral balances in premature infants fed their own mothers milk or formula. J. Pediatr. 1983; 102: 99106.

7 Valentine C, Hurst N, Schanler RJ. Hindmilk improves weight

gain in low birthweight infants fed human milk. J. Pediatr. Gastroenterol. Nutr. 1994; 18: 4747.
8 Butte NF, Garza C, Smith EO. Variability of micronutrient concentrations in human milk. Eur. J. Clin. Nutr. 1988; 42: 3459.
9 Hall B. Uniformity of human milk. Am. J. Clin. Nutr. 1979; 32:
30412.
10 Schanler RJ. Suitability of human milk for low birthweight infants. Clin. Perinatol. 1995; 22: 20723.
11 Innis SM. Human milk and formula fatty acids. J. Pediatr. 1992;
120: S5661.
12 Uauy R, Hoffman DR. Essential fatty acid requirement for normal
eye and brain development. Semin. Perinatol. 1991; 15: 44955.
13 Slusher T, Hampton R, Bode-Thomas F et al. Promoting the exclusive feeding of own mothers milk through the use of hindmilk and
increased maternal milk volume for hospitalized low birth weight
infants (<1800 grams) in Nigeria: a feasibility study. J. Hum. Lact.
2003; 19: 191198.
14 Dubowitz LM, Dubowitz W, Golgberg C. Clinical assessment
of gestational age in the newborn. J. Pediatr. 1970; 77: 110.
15 Neville MC, Keller RP, Seacat J. Studies on human lactation: 1.
Within-feed and between-breast variation in selected components of human milk. Am. J. Clin. Nutr. 1984; 40: 63546.
16 Lucas A, Gibbs JAH, Lyster RLJ, Baum JD. Creamatocrit: Simple, clinical technique for estimating fat concentration and energy value of human milk. BMJ 1978: 1: 101820.
17 Graham VC, Ingeborg CR, Willins DM et al. Postnatal growth
of infants of very low birthweight: Effects of metabolic acidosis,
of caloric intake and of calcium, sodium and phosphate supplementation. J. Pediatr. 1977; 91: 78793.
18 Morris SA, Robert WW. Acid-base equilibrium in normal infants. Pediatrics 1966; 37: 72832.
19 Okafor HU. Fluid and electrolyte therapy. In: Azubuike JC,
Nkangineme KEO (eds). Paediatrics and Child Health in a
Tropical Region, 1st edn. African Educational Services, Owerri,
1999; 14654.
20 Spencer SA, Hendrickse W, Roberton D, Hull D. Energy intake
and weight gain of very low birthweight babies fed raw expressed breastmilk. BMJ 1982; 46: 91225.
21 Dorea JG, Horner MR, Bezerra VLV, Campante ML. Variation
in major constituents of fore and hindmilk of Brazilian women.
J. Trop. Pediatr. 1982; 28: 3035.