Circadian Variations in Fat Content of Expressed Breast
Milk from Mothers of Preterm Infants
Ronit Lubetzky, MD, Yoav Littner, MD, Francis B. Mimouni, MD, FACN, Shaul Dollberg, MD, FACN, and
Dror Mandel, MD, MHA
Department Neonatology, Lis Maternity Hospital (R.L., Y.L., F.B.M., S.D., D.M.), Department of Pediatrics, Dana Children’s
Hospital (R.L.), Tel Aviv Sourasky Medical Center, the Sackler Faculty of Medicine, Tel Aviv University (R.L., Y.L., F.B.M.,
S.D., D.M.), Tel Aviv, ISRAEL
Key words: breast milk, preterm infants, fat content, circadian variations
Background: Little is known about circadian variations of the fat content in expressed human milk by
mothers of preterm infants.
Objective: To test the hypothesis that the fat content of expressed preterm human milk is higher in samples
expressed in the evening (i.e. after 3 daily meals) than in the morning (after a night-long fast).
Methods: We collected samples of expressed human milk obtained from 39 mothers of hospitalized growing
preterm infants aged 7–14 days, with a gestational age at birth ranging from 26 to 33 weeks, who routinely
expressed all their milk every 3 hours, during the day time, just before bed time, and as soon as they woke up,
using a commercial breast pump (Medela AG, Baar, Switzerland). One sample was obtained from the first
morning expression (between 0600 and 0900) and the second from the evening expression (between 2100 and
2400). The entire quantity of expressed milk was collected, mixed and measured in a capillary tube after
centrifugation at 9000 rpm for 5 minutes. Creamatocrits (CMT) were performed in duplicates. Each sample was
read independently by 2 investigators who were not aware of the origin and time of sampling and the results were
averaged. Results are expressed as mean ? SD, and analyses were by paired t-test and regression analysis.
Results: CMT was significantly higher in evening than in morning samples (7.9 ? 2.9% vs. 6.6 ? 2.8%, P ?
0.005). Neither CMT nor the morning-evening difference in CMT correlated with gestational age or birthweight.
The morning CMT correlated significantly with the evening CMT (P ? 0.001).
Conclusions: There are significant circadian variations in the CMT of expressed preterm human milk. We
speculate that if higher caloric content of expressed human milk is needed in a specific preterm infant, evening
sample, if available, should be used preferentially.
study found lower concentrations of fat in morning samples
than during the rest of the day .
We thus designed the following prospective study in order
Little is known about circadian variations of fat content
to examine morning and evening samples of human milk ex-
of expressed preterm human milk. Few studies have been
pressed by mothers of preterm infants. Specifically, we tested
conducted and have led to opposite results [1,2]. One study
the hypothesis that the fat content of expressed human milk is
found that circadian variations of fat content exist, but
higher in samples expressed in the evening (i.e. after 3 daily
without any given pattern of cyclicity , while another
meals) than in the morning (after a night-long fast).
Abbreviation: CMT ? creamatocrit.
Address correspondence to: Shaul Dollberg, MD, Department of Neonatology, Lis Maternity Hospital, Tel Aviv-Sourasky Medical Center, 6 Weizman Street, Tel Aviv,
64239, ISRAEL. E-mail: email@example.com
This study was presented, in part, at the PAS annual meeting, San-Francisco, California, 2004, May 1– 4.
Journal of the American College of Nutrition, Vol. 25, No. 2, 151–154 (2006)
Published by the American College of Nutrition
Circadian Variations in Fat Content of Breast Milk
MATERIALS AND METHODS
Table 1. Maternal and Neonatal Characteristics of the Study
Mean ? SD
We collected samples of expressed human milk obtained
Maternal age (years)
28.7 ? 4.1
from 39 mothers of hospitalized growing preterm infants aged
Pre-pregnancy weight (kg)
59.5 ? 10.2
7–14 days, with a gestational age at birth ranging from 26 to 33
Pre-pregnancy BMI* (kg/m2)
22.26 ? 4.12
weeks, who routinely expressed all their milk every 3 hours
Gestational age (weeks)
30 ? 2
1381 ? 489
during the day time, just before bed time, and as soon as they
1 minute Apgar score
7.4 ? 1.8
woke up, using a commercial breast pump (Medela AG, Baar,
5 minute Apgar score
8.5 ? 1.5
Switzerland). One sample was obtained from the first morning
* BMI ? body mass index.
expression (between 0600 and 0900) and the second one from
an evening expression (between 2100 and 2400). Each mother
CMT correlated significantly with the evening CMT (R2 ?
in the study contributed only one morning and one evening
0.302, P ? 0.001) (Fig. 1).
sample (obtained on the same day). The entire sample was
collected and mixed. The fat content was measured using the
creamatocrit (CMT) method, as previously described [3– 8].
Briefly, 75 ?L aliquots were filled into glass capillary tubes,
which were sealed at one end and spun in a hematocrit centri-
We have demonstrated that in human milk expressed by
fuge for 5 minutes at 9000 rpm. The creamatocrit was read to
mothers of preterm infants, similar to what has been observed
the nearest 0.5 mm and expressed as a percentage of the length
in term infants , there are significant circadian variations in
of the milk column in the tube. Each reading was performed in
CMT, which is consistently lower in morning than in evening
duplicate, in a blinded manner, by two investigators who were
not aware of the origin and time of sampling. The results of the
We believe that the differences observed between morning
independent readings by the 2 investigators were averaged.
and evening values of CMT reflect differences in fat content.
This method is extremely precise and reproducible with a
Indeed, Lucas et al , as well as Wang et al  have shown
coefficient of variation of less than 1%. It highly correlates with
that the CMT method correlates highly and linearly both with
total lipid biochemical measurements , making it an “accu-
total fat content and energy content of milk. Thus, we may
rate inexpensive and useful technique for estimating lipid con-
assume that morning human milk expressed in such a fashion is
centration” of expressed breast milk .
lower in fat and energy than evening milk.
The reason, or reasons for these circadian variations have
not been fully understood to date. It is possible that the circa-
dian variations in human fat content may have been influenced
Results are expressed as mean ? SD, and circadian differ-
by the degree of breast fullness  together with, or rather
ences were tested by paired t-test. Regression analysis was used
than circadian dietary habits. Indeed, total fat content of milk is
to study the correlation between CMT and gestational age or
lower in foremilk than in hind milk  and theoretically it is
birthweight, and to study the potential correlation between
possible that the breast of the women in our study were fuller
morning and evening CMT.
in the morning (after more than 3 hours without expressing)
than in the evening after the whole day of pumping every 3
hours. We must point out to the fact that Insull et al, demon-
strated that the volume of milk obtained by pumping and or
manual expression is surprisingly higher (thus more complete
Demographic and maternal characteristics of the partici-
or less incomplete) than when the breasts are nursed by the
pants in this study are presented in Table 1. Mothers were
baby . Daly et al, demonstrated that changes in the rate of
healthy, had not suffered from hypertensive disorder of preg-
milk synthesis were not directly associated with the frequency
nancy or of diabetes, and most were omnivorous, to the excep-
of breastfeeds but rather with the degree of emptying of the
tion of 2 mothers who were lacto-ovo-vegetarians.
breast . Nevertheless and regardless of the specific expla-
On average, CMT was significantly higher in evening than
nation, the fact remains that morning samples are lower in fat
in morning samples (7.9 ? 2.9% vs. 6.6 ? 2.8%, P ? 0.005).
than evening samples.
In fact, in 32 out of 39 samples (82%), evening CMT exceeded
We further speculate that dietary habits may be responsible
morning values, while in 2 samples values were identical and in
for this phenomenon. Indeed, morning samples are obtained
5 samples CMT morning values exceeded evening values.
most probably after a nightlong fast, while evening samples are
Neither CMT nor the morning-evening difference in CMT
expressed after several daytime meals. It is well known that
correlated with gestational age or birthweight. The morning
dietary fat is highly influential upon the quality of fat content in
VOL. 25, NO. 2
Circadian Variations in Fat Content of Breast Milk
Fig. 1. Correlation of morning creamatocrit (%) and evening creamatocrit (%).
human milk. For instance, fatty acid distribution in human milk
2. Weber A, Loui A, Jochum F, Buhrer C, Obladen M: Breast milk
is strikingly different in vegetarian or macrobiotic lactating
from mothers of very low birthweight infants: variability in fat and
mothers and omnivorous mothers . Insull et al, have shown
protein content. Acta Paediatr 90:772–775, 2001.
that daily fat milk production may be affected by dietary
3. Fleet IR, Linzell JL: A rapid method of estimating fat in very small
manipulations for a period of 4 days, in that it correlated with
quantities of milk. J Physiol 175:14–17, 1964.
high energy and fat intake by the mother . Whether altered
4. Lucas A, Gibbs JA, Lyster RL, Baum JD: Creamatocrit: simple
diet indeed influenced fat content between morning and
clinical technique for estimating fat concentration and energy value
of human milk. Br Med J 1:1018–1020, 1978.
evening samples is currently not known. In our study, morning
5. Lemons JA, Schreiner RL, Gresham EL: Simple method for de-
and evening CMTs were highly correlated, which might indi-
termining the caloric and fat content of human milk. Pediatrics
cate that a given dietary intake in a given mother is influential
on fat content. Moreover, limited studies with supplementary
6. Lonnerdal B, Smith C, Keen CL: Analysis of breast milk: current
feeding of poorly nourished lactating women have suggested
methodologies and future needs. J Pediatr Gastroenterol Nutr
the feasibility of improvement in volume output and in nutri-
tional quality of breast milk .
7. Jensen RG, Clark RM: Methods of lipid analysis: J Pediatr Gas-
We speculate that if a higher caloric content is needed in a
troenterol Nutr 3:296–299, 1984.
specific preterm infant fed his mother’s expressed human milk,
8. Wang CD, Chu PS, Mellen BG, Shenai JP: Creamatocrit and the
evening samples, if available, should be used preferentially.
nutrient composition of human milk. J Perinatol 19:343–346, 1999.
However, in many cases, the volume of expressed breast milk
9. Meier PP, Engstrom JL, Murtaugh MA, Vasan U, Meier WA,
may not be sufficient to allow for preferential use of evening
Schanler RJ: Mothers’ milk feedings in the neonatal intensive care
samples. Furthermore, from the fact that morning and evening
unit: accuracy of the creamatocrit technique. J Perinatol 22:646–
CMT values were highly correlated, we also suggest that more
studies should be designed to test the hypothesis that specific
10. Cregan MD, Mitoulas LR, Hartmann PE: Milk prolactin, feed
volume and duration between feeds in women breastfeeding their
dietary manipulations might allow caloric enrichment of ex-
full-term infants over a 24 h period. Exp Physiol 87:207–214,
pressed human milk.
11. Daly SE, Di Rosso A, Owens RA, Hartmann PE: Degree of breast
emptying explains changes in the fat content, but not fatty acid
composition, of human milk. Exp Physiol 78:741–755, 1993.
12. Forsum E, Lonnerdal B: Variation in the contents of nutrients of
1. Gross SJ, Geller J, Tomarelli RM: Composition of breast milk
breast milk during one feeding. Nutr Rep Int 19:815–820, 1979.
from mothers of preterm infants. Pediatrics 68:490–493, 1981.
13. Insull W Jr, Hirsch J, James T, Ahrens EH Jr. The fatty acids of
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION
Circadian Variations in Fat Content of Breast Milk
human milk. II. Alterations produced by manipulation of caloric
16. Jelliffe DB, Jelliffe EF: The volume and composition of human
balance and exchange of dietary fats. J Clin Invest 38:443–450, 1959.
milk in poorly nourished communities. Am J Clin Nutr 31:492–
14. Daly SE, Kent JC, Owens RA, Hartmann PE: Frequency and
degree of milk removal and the short-term control of human milk
synthesis. Exp Physiol 81:861–875, 1996.
15. Specker BL, Wey HE, Miller D: Differences in fatty acid composition
of human milk in vegetarian and nonvegetarian women: long-term
Received October 9, 2004; revision accepted February 16,
effect of diet. J Pediatr Gastroenterol Nutr 6:764–768, 1987.
VOL. 25, NO. 2