The Use of Multinutrient

H u m a n Mi l k F o r t i f i e r s i n
P re t e r m In f a n t s
A Systematic Review of Unanswered Questions

Francis B. Mimouni, MDa,b, Natalie Nathan, BScb,c,

Ekhard E. Ziegler, MDd, Ronit Lubetzky, MDb,c,
Dror Mandel, MDb,e,*

KEYWORDS
 Human milk  Breastfeeding  Fortifier  Macronutrients

KEY POINTS
 There is little evidence that early introduction of human milk fortification compared with
late fortification affects important outcomes such as early growth.
 There is no strong evidence that human milk–based fortifiers in otherwise exclusively hu-
man milk–fed preterm infants affect important outcomes.
 There is limited evidence that a bovine fortifier used with a combination of human milk and
bovine-based formula places the infant at a higher risk of necrotizing enterocolitis.
 There is a definite need for additional studies, incorporating also long-term outcomes, to
determine whether or not the use of human milk–based fortifiers improves outcomes.

INTRODUCTION

It has long been known that very low birth weight (VLBW) preterm infants fed exclu-
sively breast milk cannot match intrauterine growth patterns and may end up with
extrauterine growth restriction.1,2 Efforts have been made to develop liquid or powder
multinutrient products for the fortification of human breast milk.3 These fortifiers

The authors have nothing to disclose.

a
Department of Neonatology, Shaare Zedek Medical Center, 12 Shmuel Bait Street, Jerusalem
913102, Israel; b Sackler Faculty of Medicine, Tel Aviv Sourasky Medical Center, Tel Aviv Univer-
sity, Tel Aviv, Israel; c Department of Pediatrics, Dana Dwek Children’s Hospital, Tel Aviv Soura-
sky Medical Center, 6 Weizman Street, Tel Aviv 6423906, Israel; d Department of Pediatrics,
University of Iowa, Iowa City, IA, USA; e Department of Neonatology, Dana Dwek Children’s
Hospital, Tel Aviv Sourasky Medical Center, 6 Weizman Street, Tel Aviv 6423906, Israel
* Corresponding author. Department of Neonatology, Dana Dwek Children’s Hospital, Tel Aviv
Sourasky Medical Center, 6 Weizman Street, Tel Aviv 6423906, Israel.
E-mail address: dmandel@post.tau.ac.il

Clin Perinatol 44 (2017) 173–178

http://dx.doi.org/10.1016/j.clp.2016.11.011 perinatology.theclinics.com
0095-5108/17/ª 2016 Elsevier Inc. All rights reserved.
174 Mimouni et al

increase nutrient intake and are expected to improve both growth and neurodevelop-
mental outcomes.3 A recent systematic review within the Cochrane collaborative proj-
ect aimed to determine whether multinutrient fortification of human breast milk
improves important growth and developmental outcomes as compared with unforti-
fied breast milk in preterm infants without increasing the risk of adverse effects,
such as feeding intolerance or necrotizing enterocolitis (NEC).4 This systematic review
identified 14 randomized trials in which a total of 1071 infants participated. It
concluded that individual trials were generally small and had weak methodology.
Nevertheless, meta-analyses led to low-quality evidence that multinutrient fortification
of breast milk increases in-hospital rates of growth by a mean daily weight gain of
1.81 g/kg (with a 95% confidence interval [CI] 1.23–2.40), by a mean weekly length
gain of 0.12 cm (95% CI 0.07–0.17), and by a mean weekly head circumference
gain of 0.08 cm/wk (95% CI 0.04–0.12). The meta-analyses did not show a positive ef-
fect of fortification on developmental outcomes. There was also low-quality evidence
that fortification did not increase the risk of NEC in preterm infants with a typical rela-
tive risk (RR) 1.57 (95% CI 0.76–3.23). The investigators of this Cochrane review
concluded that multinutrient fortified breast milk compared with unfortified breast
milk does not significantly affect important outcomes, but that it leads to a slight in-
crease of in-hospital growth rates. As often found in the conclusion of Neonatal
Cochrane Systematic reviews,5 the investigators of this important analysis concluded
that the trials available “do not provide consistent evidence of effects on longer-term
growth or development” and that “additional trials are needed to solve this issue.”4
This excellent review was published in 2016, and there was very little chance that
we would be able to reach different conclusions because of additional, new data.
We thus elected to address other issues in our systematic review, issues that were
purposely not addressed in the Cochrane review.4 We specifically elected to deter-
mine whether studies (1) answered the question of early versus late introduction of for-
tifiers with regard to growth and/or other outcomes; and (2) had compared the
efficacy/adverse effects of human milk–based fortifiers (HBF) with that of bovine for-
tifiers (BF) in otherwise exclusively human milk–fed infants.

MATERIALS AND METHODS

We conducted this systematic review in August 2016. We included only studies

reporting the use of multinutrient human milk fortifiers. One author (NN) searched
MEDLINE, EMBASE, and Google Scholar using the following key words: human
milk, human milk fortifier, premature infant, preterm infant, human milk fortification.
We also examined the references in studies identified as potentially relevant. Four au-
thors (FB, NN, DM, and RL) screened titles and abstracts of all records identified by
the search and coded records as “order” or “exclude.” We then assessed all records
coded as “order” and made the final decision about which records to order as full-text
articles. We read the full texts to assess each article’s suitability for inclusion on the
basis of prespecified inclusion and exclusion criteria. Then the data were extracted
independently by using a data collection form to aid extraction of information on
design, methods, and participants from each included study. We assessed the quality
of evidence at the outcome level using the Grading of Recommendations Assessment,
Development, and Evaluation (GRADE) (http://www.gradeworkinggroup.org/)
approach. Disagreements were discussed until a consensus was reached. If data
from a given article were insufficient, the report was excluded from analysis. For the
purpose of potential meta-analyses, we aimed to retain only articles that had studied
the question of early versus late introduction of fortifiers and studies that compared
 Human Milk Fortifier 175

the efficacy/adverse effects of HBF versus BF in otherwise exclusively human milk–fed

preterm infants.

RESULTS

Our search allowed us to initially retrieve 2471 articles. The flow chart of Fig. 1 depicts
the retrieval and selection processes that were conducted to answer the 2 questions
(see Fig. 1).
To answer the first question (early vs late introduction), only 5 trials were identi-
fied.6–9 The first one, by Tillman and colleagues,6 published in 2012, addressed the
issue in a retrospective manner, as a pre-post comparison, and involved 53 early-
fortification infants (from day 1 of feeding), as well as 42 delayed-fortification infants
(when feedings reached 50–80 mL/kg per day). The inherent confounders associated
with the retrospective study design did not allow this study to be included in a

Comparisons were not

Theoretical model (1)

Fig. 1. Flow diagram of the analysis process.

176 Mimouni et al

meta-analysis. Nevertheless, the study did not find any differences in weight at
34 weeks postmenstrual age (PMA), but stated that the delayed-fortification group
had a “higher incidence of elevated alkaline phosphatase.” The study did not find sig-
nificant differences in outcomes, such as feeding intolerance or NEC, but was under-
powered to reach such conclusions in a definitive manner. The second study, by Maas
and colleagues,7 also addressed the issue in a retrospective manner, as a pre-post,
comparison of nonconsecutive periods, and involved 206 preterm infants. Analyses
did not focus on early versus late fortification, but rather on percentage of human
milk in the total feeds. This study also was impossible to include in a meta-analysis,
as no conclusion could be drawn from it with regard to the timing of fortification.
The third article was retrieved from a non-Medline recorded article (Iranian Journal
of Pediatrics) and its full text in the English language could not be obtained.8 In its En-
glish abstract, the article stated that no anthropometric differences and no differences
in adverse outcomes were found between the early-fortification and the late-
fortification groups, but the abstract did not define what was meant by early or late.
Its sample size was also relatively small (80 infants randomized) and the study is prob-
ably underpowered to detect small differences in effects or side effects. Another
study, by Shah and colleagues,9 consisted of a randomized controlled trial of early
(beginning at an enteral intake of 20 mL/kg per day) versus late fortification (beginning
at an enteral intake of 100 mL/kg per day). A population of 100 VLBW (<1500 g) infants
was randomized to early (n 5 50) and late (n 5 50) groups. As expected, there were
small but significant differences in protein intake between groups that persisted until
week 4 of the study. In spite of that, no significant differences were found in anthropo-
metric measurements (head circumference, length, weight, or weight velocity) or
feeding intolerance, NEC, bronchopulmonary dysplasia (BPD), patent ductus arterio-
sus (PDA), or sepsis, although the study did not report post hoc type 2 errors, and was
probably underpowered to detect small differences in effects as well as adverse ef-
fects. In a study by Sullivan and colleagues,10 2 subgroups of infants were randomized
to early (40 mL/kg per day) versus late (100 mL/kg per day) introduction of an HBF, and
the study also did not find differences in growth, feeding intolerance, or NEC.10 It was
stated only that the 2 groups did not differ between themselves, and quantitative com-
parisons were not reported. Nevertheless, and because of their heterogeneity, we felt
that the study by Shah and colleagues9 and the study by Sullivan and colleagues10
could not be combined in a meta-analysis.
We therefore concluded from the first part of this systematic review that the limited
available data do not provide strong evidence that early introduction of human milk
fortification, compared late fortification, affects important outcomes, except that it
leads to slightly increased initial protein intake.
To answer the second objective (HBF vs cow milk–based fortifier in otherwise exclu-
sively human milk–fed preterm infants), we identified 5 articles.10–14 Chronologically,
the first report by Boehm and colleagues11 compared a bovine-based commercial
fortifier (at a concentration of 3 g per 100 mL fresh human milk) with powdered
freeze-dried human milk plus minerals (at a concentration of 8 g per 100 mL fresh hu-
man milk). The study was conducted in a small number of infants (24 male VLBW in-
fants) who were randomized to the BF group (n 5 13) and the HBF group (n 5 11). The
study was very short in duration (14 days), and was not able (maybe in part because of
its duration) to determine significant differences in weight or length gain. The HBF in
that study seemed to have been custom-produced by the investigators and is not
available commercially. This study was a balance study that did not find significant
differences in nitrogen or fat intake, excretion, or retention and therefore was not
retainable for a meta-analysis.
 Human Milk Fortifier 177

The second study, by Sullivan and colleagues,10 recruited 207 infants fed their own
mother’s milk and randomized to 1 of 3 groups: one group received HBF when enteral
intake was 40 mL/kg per day, the second one received HBF when enteral intake was
100 mL/kg per day, and the third one received BF when enteral intake was 100 mL/kg
weight per day. The first 2 groups received donor human milk whenever their own
mother’s milk was not available, whereas the third group received a cow milk–based
formula when mother’s milk was not available. The 3 groups did not differ in weight,
length, or head circumference patterns. When the outcomes of NEC and late-onset
sepsis (LOS) were combined, there were very similar rates among the 3 groups. How-
ever, when only NEC was considered, both HBF groups had significantly lower rates,
and all but 1 case of surgical NEC was found in the bovine preterm formula combined
with human milk and bovine fortifier group. It must be noted that LOS in all 3 groups
did not differ significantly (P 5 .3) but the lowest numbers were surprisingly in the
bovine group, which had “masked” the effect on NEC when NEC and LOS were
analyzed together. The power of the LOS analyses was not provided, and it is unclear
whether a larger number of participants might have allowed for reaching statistical
significance.
A reanalysis of the data of Sullivan and colleagues10 (Ghandehari and colleagues12)
was later published and also showed that total parenteral nutrition days were reduced
in the HBF groups. Moreover, using the analysis of Sullivan and colleagues,10 Ganapathy
and colleagues13 developed a theoretic model of costs of NEC and cost-effectiveness of
exclusively human milk products in feeding extremely premature infants and concluded
that a 100% human milk–based diet fortified with HBF may result in net savings on med-
ical care resources by preventing NEC. Nevertheless, the study by Sullivan and col-
leagues,10 as important as it is, does not allow to answer the question of superiority of
an HBF over a BF in otherwise exclusively human milk–fed preterm infants. Because
the group fed BF received a cow milk–based formula rather than donor human milk,
the difference in feedings may have significantly affected the results. The last study,
by Cristofalo and colleagues,14 is a randomized trial of infants fed exclusively human
milk and human milk–based products compared with infants fed formula and therefore
could not be included in the analyses.
Thus, at this point, we conclude that there are limited data available and that the
data do not provide strong evidence that HBF in otherwise exclusively human milk–
fed preterm infants affects important outcomes. There is limited evidence that use
of bovine fortifier with a combination of human milk and bovine-based preterm formula
places the infant at a higher risk of NEC than use of HBF with exclusively human milk,
which confirms previous observations on higher NEC risks in infants fed preterm for-
mula as compared with human milk–fed preterm infants. However, the study design of
this trial does not allow the conclusion that the use of human milk–based fortifier would
reduce NEC risk as compared with bovine-based fortifier. There is a definite need for
additional studies incorporating long-term outcomes to determine in a clean head-to-
head comparison whether or not the use of human milk–based fortifiers might improve
outcomes and reduce NEC and other adverse outcomes.

REFERENCES

1. Kashyap S, Schulze KF, Forsyth M, et al. Growth, nutrient retention, and metabolic
response of low-birth-weight infants fed supplemented and unsupplemented pre-
term human milk. Am J Clin Nutr 1990;52:254–62.
2. Ehrenkranz RA, Younes N, Lemons JA, et al. Longitudinal growth of hospitalized
very low birth weight infants. Pediatrics 1999;104:280–9.
178 Mimouni et al

3. Ziegler EE. Human milk and human milk fortifiers. In: Koletzko B, Poindexter B,
Uauy R, editors. Nutritional care of preterm infants, vol. 110. Basel (Switzerland):
Karger Publishers; 2014. p. 215–27.
4. Brown JV, Embleton ND, Harding JE, et al. Multi-nutrient fortification of human milk
for preterm infants [review]. Cochrane Database Syst Rev 2016;(5):CD000343.
5. Mandel D, Littner Y, Mimouni FB, et al. Conclusiveness of the Cochrane neonatal
reviews: a systematic analysis. Acta Paediatr 2006;95:1209–12.
6. Tillman S, Brandon DH, Silva SG. Evaluation of human milk fortification from the
time of the first feeding: effects on infants of less than 31 weeks gestational age.
J Perinatol 2012;32:525–31.
7. Maas C, Wiechers C, Bernhard W, et al. Early feeding of fortified breast milk and
in-hospital-growth in very premature infants: a retrospective cohort analysis. BMC
Pediatr 2013;13:178.
8. Sajjadian N, Alizadeh TP, Asgharyan FM, et al. A comparison between early and
late breast milk fortification in preterm infants: a clinical trial study. Iran J Pediatr
2014;24:45.
9. Shah SD, Dereddy N, Jones TL, et al. Early versus delayed human milk fortifica-
tion in very low birth weight infants—a randomized controlled trial. J Pediatr 2016;
174:126–31.
10. Sullivan S, Schanler RJ, Kim JH, et al. An exclusively human milk-based diet is
associated with a lower rate of necrotizing enterocolitis than a diet of human
milk and bovine milk-based products. J Pediatr 2010;156:562–7.
11. Boehm G, Müller DM, Senger H, et al. Nitrogen and fat balances in very low birth
weight infants fed human milk fortified with human milk or bovine milk protein. Eur
J Pediatr 1993;152:236–9.
12. Ghandehari H, Lee ML, Rechtman DJ, H2MF Study Group. An exclusive human
milk-based diet in extremely premature infants reduces the probability of re-
maining on total parenteral nutrition: a reanalysis of the data. BMC Res Notes
2012;5:188.
13. Ganapathy V, Hay JW, Kim JH. Costs of necrotizing enterocolitis and cost-
effectiveness of exclusively human milk-based products in feeding extremely
premature infants. Breastfeed Med 2012;7:29–37.
14. Cristofalo EA, Schanler RJ, Blanco CL, et al. Randomized trial of exclusive human
milk versus preterm formula diets in extremely premature infants. J Pediatr 2013;
163:1592–5.