1. Objectives

To evaluate whether multi-nutrient fortified human breast milk compared with unfortified breast milk improves growth and development in preterm infants without increasing the risk of adverse effects such as necrotising enterocolitis and feed intolerance

2. How studies were identified

The following databases were searched in February 2016:

• CENTRAL (The Cochrane Library 2016, Issue 2)
• MEDLINE
• EMBASE
• Maternity and Infant Care
• CINAHL
• ClinicalTrials.gov
• Current Controlled Trials

Reference lists and conference proceedings were also searched

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials, including cluster-randomized and quasi-randomized trials

3.2 Study participants

Preterm (<37 weeks’ gestation) and low birth weight (<2500 g) infants receiving enteral human breast milk

3.3 Interventions

Human breast milk (expressed maternal, donor or both) fortified with energy (carbohydrate or fat) and protein in comparison with human breast milk without added energy or protein, with a minimum follow-up of two weeks

(Multi-nutrient fortifiers may have been animal or human milk-based, and both fortified and control milk may have contained additional micronutrient fortificants, such as iron)

(Infants in treatment and control groups should have received the same volume of breast milk and similar care)

3.4 Primary outcomes

Growth outcomes

• Weight
• Length
• Head circumference
• Skinfold thickness
• Body mass index and other measures of body composition
• Growth restriction (<10^th centile for the index population distribution of weight, length, or head circumference)

Neurodevelopmental outcomes (assessed >12 months post-term)

• Neurological evaluations
• Developmental scores
• Classifications of disability, including auditory and visual disability
• Neurodevelopmental impairment, defined as ≥1 of the following: non-ambulant cerebral palsy, developmental quotient >2 standard deviations below the population mean, blindness (visual acuity <6/60) or deafness (hearing impairment requiring amplification)

Secondary outcomes included duration of hospitalization, feed intolerance resulting in cessation/reduction of enteral feeding, necrotising enterocolitis, bone mineralization (as measured by dual energy X-ray absorptiometry, serum alkaline phosphatase concentrations, clinical or radiological evidence of rickets), and long-term metabolic or cardiovascular health (as measured by insulin resistance, obesity, diabetes, hypertension)

4. Main results

4.1 Included studies

Fourteen randomized controlled trials, enrolling 1071 infants, were included in this review

• Eight trials included only very low birth weight (<1500 g) or very preterm (<32 weeks’ gestation) infants, and the remaining trials specified maximum birth weights for study entry ranging between 1600 and 2000 g
• Across all included trials the average gestational age was 30 weeks, all infants must have been able to tolerate enteral feeds, and all mothers must have been able to provide expressed breast milk, although in seven trials human donor milk was used to supplement intake
• Eight trials used commercially available cow’s milk-based powdered fortifier containing protein, fat, carbohydrate, minerals and electrolytes. Two trials mixed human milk with an equal volume of preterm formula, and three trials did not specify the type or brand of fortifier used
• Additional micronutrient fortification was also added to control infants’ feeds in nine trials

4.2 Study settings

• Canada, Denmark, India (2 trials), Italy, Oman, South Africa (2 trials), Sweden, the United Kingdom of Great Britain and Northern Ireland (2 trials), and the United States of America (3 trials)
• Included trials were published between 1986 and 2012
• All trials were conducted in hospital settings

4.3 Study settings

How the data were analysed
The effects of human milk fortified with energy and protein were compared to those of human breast milk without added energy or protein in preterm infants. For dichotomous data, risk ratios (RR) were calculated, while for continuous data, mean differences (MD) were produced. Corresponding 95% confidence intervals (CI) were also generated. Heterogeneity was deemed substantial if I² was greater than 50%. To investigate potential sources of heterogeneity, sensitivity analyses were undertaken, and the following subgroup analyses were planned:

• Very preterm (<32 weeks’ gestation) or very low birth weight (<1500 g) infants
• Fortification of donor breast milk
• Trials using human milk-based fortifier
• Trials fortifying breast milk with infant formula
• Trials conducted in low- and middle-income countries

Results
Growth
Weight gain
Overall, weight gain was statistically significantly greater by 1.81 g/kg/day among infants receiving fortified breast milk in comparison with those receiving unfortified milk (95% CI [1.23 to 2.40 g/kg/day], p<0.00001; 10 trials/635 infants). The effect was similar in subgroup analysis of two trials conducted in low- or middle-income countries (MD 1.86 g/kg/day, 95% CI [0.70 to 3.01], p<0.00001; 214 infants), and was greater in subgroup analysis of trials in very preterm or very low birth weight infants (MD 2.82 g/kg/day, 95% CI [1.83 to 3.80], p<0.00001; 5 trials/269 infants).

Length gain
Overall, length gain was improved with the use of fortified breast milk by an average of 0.12 cm/week (95% CI [0.07 to 0.17 cm/week], p<0.00001; 8 trials/555 infants). The effect size was slightly increased in subgroup analyses of trials conducted in low- or middle-income countries (MD 0.18 cm/week, 95% CI [0.10 to 0.26], p<0.00001; 1 trial/157 infants) and trials in very preterm or very low birth weight infants (MD 0.21 cm/week, 95% CI [0.14 to 0.28], p<0.0001; 3 trials/189 infants).

Change in head circumference
Overall, a significant improvement in head circumference growth was found among infants receiving fortified milk (MD 0.08 cm/week, 95% CI [0.04 to 0.12], p<0.0001; 8 trials/555 infants). Similar effects were found in subgroup analyses of trials conducted in low- or middle-income countries (MD 0.08 cm/week, 95% CI [0.02 to 0.14], p=0.01; 1 trial/157 infants) and trials in very preterm or very low birth weight infants (MD 0.11 cm/week, 95% CI [0.05 to 0.17], p<0.0001; 3 trials/189 infants).

Anthropometry at 12 to 18 months
Pooled analysis of two trials involving 270 infants demonstrated no statistically significant differences in weight (MD -0.03 kg, 95% CI [-0.31 to 0.25]), length (MD -0.19 cm, 95% CI [-0.98 to 0.60]), or head circumference (MD 0.10 cm, 95% CI [-0.37 to 0.18]) between treatment groups at 12 to 18 months of age.

Neurodevelopmental outcomes after 12 months
In one trial of 245 infants, no statistically significant difference between fortified and unfortified groups was found for the outcomes mental development index (MD 2.20 points, 95% CI [-3.35 to 7.75]) and psychomotor development index (MD 2.40 points, 95% CI [-1.90 to 6.70]) at 18 months of age.

Additional outcomes
Meta-analysis of two trials involving 210 infants found no statistically significant difference in length of hospital stay between treatment groups. In pooled analysis of five trials in which the control group did not receive mineral fortification, serum alkaline phosphatase was significantly lower among the fortified group (MD -126 IU/L, 95% CI [-191 to -62], p=0.000089; 325 infants), indicating less bone turnover. In one trial of 59 infants, bone mineral content was increased in the fortified group compared with the unfortified group (MD 12.0 mg/cm, 95% CI [6.3 to 17.7], p=0.00004).

Adverse effects
Overall, the risk of necrotising enterocolitis was not significantly increased among infants receiving fortified milk (RR 1.57, 95% CI [0.76 to 3.23], 11 trials/882 infants), and nor was the risk of feed intolerance (RR 0.90, 95% CI [0.54 to 1.49], 5 trials/255 infants). Results were not altered meaningfully in subgroup analyses of trials in low- or middle-income countries and trials in very preterm or very low birth weight infants.

5. Additional author observations*

The overall methodological quality of the included trials was poor and sample sizes were low. GRADE quality of evidence assessments rated the outcomes weight gain, length gain, and necrotising eneterocolitis as low quality, while the outcomes head circumference growth, mental development index, and psychomotor development index were rated as being of moderate quality. Most included trials were conducted in high-income countries, limiting the generalizability of the findings to resource-constrained settings.

Multi-nutrient fortification of breast milk improved rates of weight gain, length gain and head circumference growth in preterm infants. However, the effect sizes were relatively small, and although few long-term data were available for analysis, it appeared that the improvements in growth were only short-term.

Multi-nutrient fortification of breast milk is already widely established in current clinical practice, limiting the potential for future trials in this area. However, the long-term effects of feeding preterm infants with fortified breast milk are yet to be established. In addition, evaluation of the separate effects of added micronutrients is warranted.