Compared with term infants, preterm infants have limited nutrient reserves at birth. Preterm infants, especially very preterm or very low birth weight (VLBW) infants, are additionally subject to a variety of physiological and metabolic stresses that increase their nutrient needs. Recommended nutrient requirements for preterm infants are based on intrauterine growth studies and assume that the optimal rate of postnatal growth should be about the same as that of normal, uncompromised fetuses of an equivalent postmenstrual age. However, evidence exists that in practice the recommended target levels of nutrient input are rarely achieved and most very preterm or VLBW infants accumulate significant energy, protein, mineral, and other nutrient deficits during their hospital stay (Embleton 2001). Consequently, many preterm infants and most very preterm or VLBW infants are growth restricted by the time they are ready for hospital discharge (Lucas 1984; Clark 2003).
Description of the condition
Following hospital discharge, ad libitum (demand) fed preterm infants often consume more milk than term infants in order to attain some "catch up" growth (Lucas 1992a). Despite this, growth deficits can persist through childhood and adolescence (Ford 2000; Farooqi 2006; Trebar 2007; Bracewell 2008). Slow or incomplete catch up growth in preterm infants, especially of the head, is associated with a higher risk of neurodevelopmental impairment in later childhood, as well as with poorer cognitive and educational outcomes (Hack 1991; Cooke 2003). Preterm infants who have accumulated deficits in calcium and phosphate are at higher risk of poor bone mineralization, metabolic bone disease, and a reduced rate of skeletal growth compared to infants born at term (Rigo 2000).
Description of the intervention
Because slow or incomplete catch up growth is associated with prolonged growth restriction and with slower neuro-developmental progression, attention has focused on nutritional interventions that might promote growth during the putative ‘critical window’ of early infancy in the post-discharge period. Two broad strategies for nutritional interventions exist (Dusick 2003; Fewtrell 2003; Klingenberg 2011):
- multi-nutrient fortification of expressed milk for infants fed with human breast milk
- nutrient-enriched formula for formula fed infants
Another Cochrane review addresses the question of whether multi-nutrient fortification of human breast milk affects growth and development in preterm infants following hospital discharge (McCormick 2010). This review focuses on the comparison of nutrient-enriched formula milk versus standard term formula for formula fed preterm infants following hospital discharge.
- standard term formula: designed for term infants, based on the composition of mature human breast milk. The typical energy content is 66 to 68 kcal/100 ml. The concentration of protein, approximately 1.4 to 1.7 g/100 ml, and calcium and phosphate content (about 50 mg/100 ml and 30 mg/100 ml respectively) are not sufficient to provide the recommended nutrient needs for stable and growing preterm infants.
- post-discharge formula: specifically designed for preterm infants post-discharge from hospital. These are energy (about 72 to 74 kcal/100 ml) and protein (about 1.8 to 1.9 g/100 ml) enriched, and variably enriched with minerals, vitamins, and trace elements compared to standard term formula. Expert bodies and authorities recommend these formulae for preterm infants for three to twelve months post-discharge (Aggett 2006).
- preterm formula: energy-enriched (about 80 kcal/100 ml), protein-enriched (2.0 to 2.4 g/100 ml), and variably enriched with minerals, vitamins, and trace elements to support intra-uterine nutrient accretion rates. These formulae are commonly used for nutrition of preterm infants prior to hospital discharge and are not generally recommended for post-discharge feeding.
How the intervention might work
In theory, feeding preterm infants following hospital discharge with formula enriched with extra energy, protein, minerals and vitamins may be expected to promote more rapid catch up growth. However, because preterm infants fed in response to hunger and satiation cues (ad libitum or demand) adjust their volume of intake depending upon the energy-density of the formula, infants may consume less nutrient-enriched milk than standard term formula (Lucas 1992a). Consequently, infants fed ad libitum with preterm or post-discharge formula may not receive any more nutrients than infants who receive standard term formula. Feeding with nutrient-enriched formula may also be associated with disordered gastric motility and emptying (Hancock 1984; Siegel 1984). Nutrient-enriched formula may therefore be more poorly tolerated, so reducing nutrient delivery, and potentially removing any benefits for growth and development. Furthermore, concern exists that catch up growth with accelerated weight gain and crossing of body mass index (BMI) percentiles might be associated with altered fat distribution and related ‘programmed’ metabolic consequences that may increase the risk of insulin resistance and cardiovascular disease (Hack 2003; Doyle 2004; Euser 2005; Saigal 2006; Euser 2008).
Why it is important to do this review
Given the potential for post-discharge nutrition strategies to affect growth and development in preterm infants, and since uncertainty exists about the balance between the putative benefits and harms, an attempt to detect, appraise, and synthesise evidence from randomised controlled trials is needed.
To determine how feeding preterm infants following hospital discharge with nutrient-enriched formula (preterm formula or post-discharge formula) compared with standard term formula affects growth and development.
Criteria for considering studies for this review
Types of studies
Controlled trials using random or quasi-random patient allocation. Studies published as abstracts were only eligible for inclusion if assessment of study quality was possible and if other criteria for inclusion were fulfilled.
Types of participants
Preterm infants fed with formula (exclusively or as a supplement to human breast milk) following discharge from hospital. The intervention may have commenced up to one week prior to planned discharge from hospital. Trials that randomly assigned infants to nutrient-enriched formula versus standard term formula more than one week prior to hospital discharge (and then continued the intervention after hospital discharge) were not to be included in this review.
Types of interventions
- Standard term formula: energy content < 72 kcal/100 ml, and protein content < 1.7 g/100 ml.
- Post-discharge formula: energy content > 72 kcal/100 ml (but < 75 kcal/100 ml) and protein content > 1.7 g /100 ml.
- Preterm formula: energy content between > 75 kcal/100 ml and protein content > 2.0 g/100 ml).
The formulae could be fed either as sole diet or as a supplement to human breast milk. Infants in the trial groups should have received similar care other than the type of formula. For example, there should not have been any differences between groups in the prescription of target levels of volume of intake, or advice or support for demand feeding.
Types of outcome measures
- Growth: Weight, length, head growth, skinfold thickness, BMI and measures of body composition (lean/fat mass) and growth-restriction (proportion of infants who remain < 10th percentile for the index population's distribution of weight, length, or head circumference). Long-term growth and growth-restriction (proportion of infants who remain below the tenth percentile for the index population's distribution of weight, height, or head circumference).
- Neurodevelopmental outcomes assessed using validated tools at > 12 months corrected age, and classifications of disability, including non-ambulant cerebral palsy, developmental delay, auditory and visual impairment
- Cognitive and educational outcomes at > 5 years: Intelligence quotient and/or indices of educational achievement measured using a validated tool (including school examination results).
- Feed intolerance such as vomiting or diarrhoea that necessitates ceasing the study formula.
- Measures of bone mineralization such as serum alkaline phosphatase level, or bone mineral content assessed by dual energy x ray absorptiometry and clinical or radiological evidence of rickets on long term follow-up.
- Blood pressure on long term follow-up.
- Body mass index or other measures of overweight or obesity on long term follow-up.
Search methods for identification of studies
We used the standard search strategy of the Cochrane Neonatal Review Group.
We searched the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 4, 2011), MEDLINE (1966 September 2011), EMBASE (1980 to September 2011), and CINAHL (1982 to September 2011) using a combination of the following text words and MeSH terms: [Infant, Newborn OR Infant, Premature OR Infant, Low Birth Weight OR infan* OR neonat*] AND ["Infant-Nutrition"/ all subheadings OR Infant Formula OR milk OR formula OR post-discharge OR fortif* OR supplement*].
The search outputs were limited with the relevant search filters for clinical trials as recommended in the Cochrane Handbook. We did not apply any language restriction.
Searching other resources
We examined the references in studies identified as potentially relevant. We also searched the abstracts from the annual meetings of the Pediatric Academic Societies (1993 to 2011), the European Society for Pediatric Research (1995 to 2011), the UK Royal College of Paediatrics and Child Health (2000 to 2011), and the Perinatal Society of Australia and New Zealand (2000 to 2011). We considered trials reported only as abstracts to be eligible if sufficient information was available from the report, or from contact with the authors, to fulfil the inclusion criteria.
Data collection and analysis
We used the standard methods of the Cochrane Neonatal Review Group.
Selection of studies
Two review authors screened the title and abstract of all studies identified by the above search strategy. We reassessed the full text of any potentially eligible reports and excluded those studies that did not meet all of the inclusion criteria. We discussed any disagreements until consensus was achieved.
Data extraction and management
We used a data collection form to aid extraction of relevant information from each included study. Two review authors extracted the data separately. We discussed any disagreements until consensus was achieved. We asked the investigators for further information if data from the trial reports were insufficient.
Assessment of risk of bias in included studies
We used the criteria and standard methods of the Cochrane Neonatal Review Group to assess the methodological quality of any included trials. Additional information from the trial authors was requested to clarify methodology and results as necessary. We evaluated and reported the following issues in the Risk of Bias tables:
- Sequence generation: We categorised the method used to generate the allocation sequence as:
- low risk (any random process e.g. random number table; computer random number generator);
- high risk (any non random process e.g. odd or even date of birth; patient case-record number);
- Allocation concealment: We categorised the method used to conceal the allocation sequence as:
- low risk (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);
- high risk (open random allocation; unsealed or non-opaque envelopes, alternation; date of birth);
- Blinding: We assessed blinding separately for different outcomes and categorised the methods as:
- low risk, high risk or unclear for
- clinicians and caregivers and;
- outcome assessors.
- Incomplete outcome data: We described the completeness of data including attrition and exclusions from the analysis for each outcome and any reasons for attrition or exclusion where reported. We assessed whether missing data were balanced across groups or were related to outcomes. Where sufficient information was reported or supplied by the trial authors, re-included missing data in the analyses. We categorised completeness as:
- low risk (< 20% missing data);
- high risk (> 20% missing data);
Measures of treatment effect
We calculated relative risk (RR) and risk difference (RD) for dichotomous data and weighted mean difference (WMD) for continuous data, with respective 95% confidence intervals (CI). We determined the number needed to treat for benefit (NNTB) or harm (NNTH) for a statistically significant difference in the RD.
Unit of analysis issues
The unit of analysis is the participating infant in individually randomised trials and the neonatal unit for cluster randomised trials.
Assessment of heterogeneity
If more than one trial was included in a meta-analysis, we examined the treatment effects of individual trials and heterogeneity between trial results by inspecting the forest plots. We calculated the I² statistic for each analysis to quantify inconsistency across studies and describe the percentage of variability in effect estimates that may be due to heterogeneity rather than sampling error. If substantial (I² > 50%) heterogeneity was detected, we explored the possible causes (for example, differences in study design, participants, interventions, or completeness of outcome assessments) in sensitivity analyses.
Assessment of reporting biases
If more than five trials were included in a meta-analysis, we conducted a funnel plot analysis.
We used the fixed effect model in RevMan 5.1 for meta-analysis.
Subgroup analysis and investigation of heterogeneity
We pre-specified the following sub-group analyses:
- very preterm (< 32 weeks) or VLBW (< 1500 g) infants;
- infants who were small for gestational age (< 10th percentile for weight) at hospital discharge;
- infants with chronic lung disease receiving supplemental oxygen therapy at hospital discharge.
Description of studies
We identified 15 trials that fulfilled the eligibility criteria (Lucas 1992; Atkinson 1999; Lucas 2001; Cooke 2001; Carver 2001; De Curtis 2002; Agosti 2003; Litmanovitz 2004; Atkinson 2004; Peng 2004; Picaud 2005; Koo 2006; Taroni 2009; Roggero 2011a; Jeon 2011; see Characteristics of included studies).
The trials were undertaken within the last 20 years by investigators attached to perinatal centres in Europe, North America and the middle-East. In total, 1128 infants have participated in these trials (range 20 to 229).
Most trials specified a maximum birth weight as the primary eligibility criterion:
Three trials specified gestational age as an eligibility criterion:
- < 35 weeks: Koo 2006
Three trials specifically recruited participants who were small for gestational age:
- birth weight < 5th percentile: Atkinson 2004
Of the other trials, although most reports did not specify intra-uterine or postnatal growth restriction as exclusion criteria, it appears that very few participants in the trials were small for gestational age at birth or enrolment.
Generally, infants with additional problems at discharge, particularly inadequate independent oral feeding or receipt of supplemental oxygen secondary to chronic lung disease, were not eligible to participate.
POST-DISCHARGE FORMULA VERSUS STANDARD TERM FORMULA (Comparison 1)
10 trials (N = 762): Lucas 1992; Atkinson 1999; Carver 2001; Lucas 2001; De Curtis 2002; Atkinson 2004; Litmanovitz 2004; Koo 2006;Taroni 2009; Roggero 2011a.
All of the participating infants were fed ad libitum. These feeds were intended to be the principal source of milk for a range of periods post term (or post hospital discharge):
- one month:Taroni 2009
The main outcomes assessed were growth parameters (weight, length, and occipito-frontal head circumference) assessed up to 12 to 18 months corrected age. Three trials assessed neuro-developmental outcomes at 18 months using Bayley Scales of Infant Development II (Lucas 2001; Cooke 2001; Jeon 2011). One trial assessed Griffiths' Developmental Scales at six, nine and twelve months corrected age(Agosti 2003).
We excluded nine studies (Cooper 1985; Bernbaum 1989; Bhatia 1991; Friel 1993; Chan 1994; Wheeler 1996; Brunton 1998; Lapillonne 2004; Amesz 2010). The reasons for exclusion are listed in the table, Characteristics of excluded studies.
Risk of bias in included studies
The trials were of variable methodological quality (Figure 1).
|Figure 1. Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.|
In seven trials, the described method of randomisation was likely to ensure blinding of allocation (Atkinson 1999; Cooke 2001; Lucas 2001; Atkinson 2004; .Picaud 2005; Koo 2006; Cooke 2001). In the other trials, it is not clear whether allocation concealment was adequate. In one of these trials, substantial between-group differences in baseline demographic characteristics existed, most likely due to allocation bias (Jeon 2011). This trial originally randomised participants to one of three intervention groups. We elected to discard data from the group where infants' characteristics were statistically significantly different from other groups at enrolment.
Most of the trials blinded families to the type of milk that the infant received. In two trials, the families are likely to have been aware which type of milk their infant had been allocated to receive (Agosti 2003; Peng 2004; Jeon 2011). It is unclear whether blinding was satisfactory in another three trials (Litmanovitz 2004; Taroni 2009; Roggero 2011a).
Most of the trials blinded outcomes assessors and investigators to the type of milk that the infant received but in four of the trials it is unclear if this was satisfactory (Litmanovitz 2004;Taroni 2009; Jeon 2011; Roggero 2011a), and in one trial physicians were unblinded (Peng 2004).
Incomplete outcome data
Ten of the trials achieved complete or near-complete (> 80%) follow-up assessment (Lucas 1992; Lucas 2001; Cooke 2001; De Curtis 2002; Atkinson 2004; Litmanovitz 2004; Peng 2004; Taroni 2009; Roggero 2011a; Jeon 2011). In two of the other trials, 75% of infants underwent outcomes assessments at latest follow-up (Atkinson 1999; Koo 2006). In another two trials, < 50% of infants completed the planned 12 months follow-up assessment (Carver 2001; Agosti 2003). In Picaud 2005, loss to follow-up by 12 months in the control group was substantial (35%) and greater than that in the intervention group (9%).
Effects of interventions
POST-DISCHARGE FORMULA VERSUS STANDARD TERM FORMULA (COMPARISON 1)
Growth (Outcomes 1.1- 1.4)
Lucas 1992 did not detect any statistically significantly differences in weight, length or head circumference at the end of the intervention and follow-up period (nine months corrected age).
Atkinson 1999 reported that infants who received post-discharge formula were statistically significantly heavier at six, nine and twelve months corrected age. There were not any statistically significant differences in length or head circumference.
Carver 2001 did not detect any statistically significant differences in weight, length, or head circumference at six and twelve months corrected age. There was substantial loss to follow-up during the trial and since published report does not state how many infants were assessed at the various time points the data could not be used to calculate mean differences.
Lucas 2001 reported that at completion of the intervention period (nine months corrected age), weight and length were statistically significantly greater in infants who received post-discharge formula but that there was not a statistically significant difference in head circumference. At 18 months, there were not any statistically significant differences in weight or head circumference. The group of infants who received post-discharge formula remained statistically significantly longer on average than the control group [mean difference 9.0 (95% CI 0.3 to 17.7) mm].
De Curtis 2002 did not find any statistically significant differences in the rate of gain of weight, length, or head circumference during the two months trial period.
Atkinson 2004 did not find any statistically significant differences in the rate of gain of weight, length, or head circumference up to 12 months corrected age (growth data reported as z scores).
Litmanovitz 2004 did not find any statistically significant differences in the weight, length, or head circumference at six months corrected age.
Koo 2006 reported that the mean weight, head circumference, and length was lower in the nutrient-enriched formula group at six, nine and twelve months after hospital discharge.
Taroni 2009 did not find any statistically significant differences in weight, length, or head circumference at one month corrected age.
Roggero 2011a did not find any statistically significant differences in weight, length, or head circumference at six months corrected age.
Meta-analyses of growth data
- Weight (Outcome 1.2; Figure 2): Meta-analyses did not detect any statistically significant differences in weight at three to four and six months corrected age. At nine months, meta-analysis of data from four trials (Lucas 1992; Atkinson 1999; Koo 2006; Lucas 2001) indicated that infants in the post-discharge formula group were heavier [WMD: 244 (95% CI 17 to 471) g]. At follow-up at 12 months, there was not a statistically significant difference.
Figure 2. Forest plot of comparison: 1 POST-DISCHARGE FORMULA VERSUS STANDARD TERM FORMULA, outcome: 1.2 Weight (g).
- Length (Outcome 1.3; Figure 3): Meta-analyses did not detect any statistically significant differences in weight at three to four and six months corrected age. At nine months, meta-analysis of data from four trials (Lucas 1992; Atkinson 1999; Koo 2006; Lucas 2001) indicated that infants in the post-discharge formula group were longer [WMD: 7.3 (95% CI 1.8 to 12.9) mm]. At follow-up at 12 months, there was not a statistically significant difference.
Figure 3. Forest plot of comparison: 1 POST-DISCHARGE FORMULA VERSUS STANDARD TERM FORMULA, outcome: 1.3 Crown heel length (mm).
- Head circumference (Outcome 1.4; Figure 4): Meta-analyses did not detect any statistically significant differences at three to four, six, nine or twelve months.
Figure 4. Forest plot of comparison: 1 POST-DISCHARGE FORMULA VERSUS STANDARD TERM FORMULA, outcome: 1.4 Head circumference (mm).
These meta-analyses all contained substantial statistical heterogeneity (I
Development (Outcome 1.5)
Lucas 2001 did not detect a statistically significant difference in the Bayley Scales Mental or Psychomotor Development Index assessed at 18 months corrected age. None of the included trials assessed later cognitive and educational outcomes.
Only one trial assessed this outcome (Lucas 1992). There was not a statistically significant difference in the mean number of vomits or possets per day. None of the participating infants ceased taking a study formula because of feed intolerance.
Bone mineralization (Outcome 1.6:)
Atkinson 2004 did not find any statistically significant differences in bone mineral content assessed at 12 months corrected age (numerical data not available).
De Curtis 2002 did not find any statistically significant differences in the bone mineral content or the bone area at the end of the two months study period.
Koo 2006 reported that at the end of the 12 months study period the infants who received nutrient-enriched formula had statistically significantly lower bone mass (measured using dual-energy X-ray absorptiometry). The data were presented in graphs and could not be extracted or obtained for calculation of mean differences.
Lucas 1992 assessed bone width and bone mineral content of the radius at nine months corrected age. The bone width was not statistically significantly different between the groups. The bone mineral content was statistically significantly higher in the group of infants who received the post-discharge formula: Mean difference 20.6 (95% CI 7.8 to 33.4) mg/cm.
Litmanovitz 2004 did not find any statistically significant differences in bone strength assessed as "bone speed of sound" measured with ultrasound or in serum levels of bone specific alkaline phosphatase at six months corrected age.
None of the trials assessed the effect of the intervention on clinical or radiological evidence of rickets.
Blood pressure on long term follow-up
Not assessed by any of the included trials.
Boby mass index on long term follow-up
Not assessed by any of the included trials.
- VLBW or very preterm infants: Two trials recruited exclusively VLBW infants (Litmanovitz 2004;Taroni 2009). As described above, the investigators did not find any statistically significant difference in the weight, length, or head circumference, or in measures of bone mineralization up to six months corrected age.
- Infants who remain small for gestational age at hospital discharge: Three trials recruited growth-restricted at birth infants (Atkinson 2004; Taroni 2009; Roggero 2011a).These trials did not detect any statistically significant effects on weight up to 12 months corrected age, but meta-analyses of data from the two trials that undertook follow-up at six months found statistically significant effects on crown heel length (WMD 8.88 [95% CI 0.94 to 16.83] mm) and head circumference (5.36 [95% CI 0.62 to 10.11] mm) (Atkinson 2004; Roggero 2011a).
- Infants with chronic lung disease requiring home supplemental oxygen therapy: None of the trials recruited exclusively infants with chronic lung disease. Subgroup data were not available.
PRETERM FORMULA VERSUS STANDARD TERM FORMULA (COMPARISON 2)
Growth (Outcomes 2.1- 2.4)
Cooke 2001 did not find a statistically significant difference in rate of weight gain during the trial period. These data were presented in graphs only and were not able to be extracted to allow calculation of the mean difference. At 18 months corrected age, the nutrient-enriched formula group was statistically significantly heavier than the control group [mean difference: 500 (95% CI 26 to 974) g], but there were not any statistically significant differences in length or head circumference.
Agosti 2003 did not find any statistically significant differences in mean weight, length, or head circumference at four, six and twelve months after hospital discharge.
Peng 2004 did not find any statistically significant differences in mean weight, length, or head circumference at monthly intervals up to six months corrected age.
Picaud 2005 did not find any statistically significant differences in the rate of gain of weight, length, or head circumference during the initial four months trial period. There were not any statistically significant differences in weight, length or head circumference between the groups at four months. At 12 months post-discharge, infants in the preterm formula group were heavier [mean difference:1007 (95% CI 211 to 1803) g] (Outcome 2.2), longer [mean difference: 27 (95% CI 2 to 52) mm] (Outcome 2.3), and had larger head circumferences [mean difference:12 (95% CI 0.2 to 24) mm] (Outcome 2.4) than control infants. However, loss to follow-up by 12 months in the control group was substantial (35%) and greater than that in the intervention group (9%).
Jeon 2011 did not find any statistically significant differences in mean weight, length, or head circumference at three, twelve and eighteen months after hospital discharge.
Meta-analyses of growth data
- Weight (Outcome 2.2; Figure 5): Weight (Outcome 2.2; Figure 5): Meta-analysis of data from four trials (Cooke 2001; Agosti 2003; Picaud 2005; Jeon 2011) found a statistically significant higher weight in the preterm formula group at 12 months corrected age [WMD: 540 (95% CI 255 to 824) g]. Meta-analysis of data from two trials (Cooke 2001; Jeon 2011) found a statistically significant higher weight in the preterm formula group at 18 months [WMD: 491 (95% CI 142 to 839) g] (Outcome 2.2).
Figure 5. Forest plot of comparison: 2 PRETERM FORMULA VERSUS STANDARD TERM FORMULA, outcome: 2.2 Weight (g).
- Length (Outcome 2.3; Figure 6): Meta-analysis of data from three trials (Agosti 2003; Picaud 2005; Jeon 2011) did not detect a statistically significant difference at 12 months corrected age [WMD: 5.1 (95% CI -4.2 to 14.5) mm]. Meta-analysis of data from two trials (Cooke 2001; Jeon 2011) found a statistically significant higher crown heel length in the preterm formula group at 18 months [WMD:11 (95% CI 2 to 20) mm]
Figure 6. Forest plot of comparison: 2 PRETERM FORMULA VERSUS STANDARD TERM FORMULA, outcome: 2.3 Crown heel length (mm).
- Head circumference (Outcome 2.4; Figure 7): Meta-analysis of data from three trials (Agosti 2003; Jeon 2011; Picaud 2005) found a statistically significant larger head circumference in the preterm formula group at 12 months corrected age [WMD: 6.1 (95% CI 1.1 to 11.1) mm]. Meta-analysis of data from two trials (Cooke 2001; Jeon 2011) found a statistically significant larger head circumference in the preterm formula group at 18 months [WMD: 5.4 (95% CI 0.7 to 10.1) mm].
Figure 7. Forest plot of comparison: 2 PRETERM FORMULA VERSUS STANDARD TERM FORMULA, outcome: 2.4 Head circumference (mm).
Development (Outcome 2.5)
Neither Cooke 2001 or Jeon 2011, nor a meta-analysis of data from both trials detected a statistically significant difference in the Bayley Scales Mental Development Index [WMD -1.4 (95% CI -6.2 to 3.4)] or Psychomotor Development Index [WMD -1.1 (95% CI -4.2 to 1.93)]. Agosti 2003 did not detect any statistically significant differences in the Griffiths' Developmental Scale evaluations at six, nine and twelve months corrected age (numerical data not available from report or trialists).
Not assessed by any of the included trials.
Cooke 2001 assessed body composition with dual energy x-ray absorptiometry at six and twelve months corrected age. There were not any statistically significant differences in the bone area, bone mineral mass, or bone mineral density measurements between the groups. In the published report, all of these data were presented in graphs and could not be extracted for estimation of mean differences. The investigators also reported that there were not any statistically significant differences in the serum phosphorus, calcium and alkaline phosphatase levels measured at intervals during the study period (up to six months post term). These data were presented mainly in graphs and could not be extracted for estimation of mean differences.
Blood pressure on long term follow-up
Not assessed by any of the included trials.
Boby mass index on long term follow-up
Not assessed by any of the included trials.
- Infants who remain small for gestational age (less than 10th percentile for weight) at hospital discharge: Subgroup data were not available.
- Infants with chronic lung disease requiring home supplemental oxygen therapy: Subgroup data were not available.
Summary of main results
Data from ten randomised controlled trials with a total of 762 participants did not provide consistent evidence that feeding preterm infants after hospital discharge with post-discharge formula (˜74 Kcal/100 ml) versus standard term formula (˜67 Kcal/100 ml) affects growth parameters up to 12 to 18 months corrected age.
The five trials that examined the effect of feeding with preterm formula (˜80 Kcal/100 mL) versus standard term formula provided stronger evidence of an effect on growth parameters. Meta-analyses found a weighted mean difference of about 500 grams for weight, 11 mm for length, and 5 to 6 mm for head circumference at 12 to 18 months corrected age. However, it is not yet known whether any of these differences persist through later childhood.
The evidence of the effect of nutrient-enriched formula on long term development is also unclear. The only trial of post-discharge versus term formula to assess developmental outcomes did not detect a statistically significant difference in the Bayley Scales Mental or Psychomotor Development Index assessed at 18 months corrected age (Lucas 2001). However, the 95% confidence intervals for the estimates of effect are wide and do no exclude modest but potentially important effect sizes. Similarly, meta-analyses of data from two trials (Cooke 2001; Jeon 2011) did not provide evidence that feeding with preterm versus term formula affects neurodevelopmental outcomes at 18 months corrected age. There are not yet any data on longer-term cognitive and educational outcomes.
Overall completeness and applicability of evidence
Although we identified ten eligible trials that compared feeding with post-discharge formula versus term formula, these were generally small and of variable methodological quality. Quantitative synthesis was limited as only six of the trials presented data that could be included in meta-analyses of growth outcomes (Lucas 1992; Atkinson 1999; Lucas 2001; Litmanovitz 2004; Koo 2006; Roggero 2011a). Interpretation of the meta-analyses was further limited by substantial and statistically significant heterogeneity. The source of heterogeneity is not clear as these trials were of similar design (intervention given for 6 to 12 months) and methodological quality (all had satisfactory processes to ensure allocation concealment and all achieved about 70% to 80% follow-up at > 6 months corrected age). The meta-analyses of data from the five trials that compared preterm formula versus term formula were more complete and did not demonstrate statistical heterogeneity.
The explanation for the difference in the measured effect on growth parameters of post-discharge formula and preterm formula may simply be related to total nutrient content and intake. An additional factor is that whereas post-discharge formula contains about 10% more calories and 20% to 25% more protein and bone minerals than term formula, preterm formula is about 20% energy-enriched and contains 40% to 60% more protein and minerals than term formula. Since ad libitum fed infants regulate their volume of milk intake relative to its calorie-density, infants in the comparison groups may have received similar total energy intakes. However, infants fed with post-discharge formula would still have received about 10% more protein and minerals than term formula fed infants, whereas infants fed with preterm formula would have received up to about 25% more protein and minerals than term formula-fed infants. It is possible that the protein and mineral intake (per unit of energy) is the key factor in determining catch-up growth rates, and specifically lean and skeletal growth, in this population of infants.
The applicability of the currently available data is limited by the short duration of follow-up in the trials. None of the trials planned or undertook any assessment of growth or development beyond 12 to 18 months corrected age and some of the trials have only reported growth outcomes up to six months. Similarly, none of the trials have reported data related to possible adverse metabolic consequences of nutrient-supplementation in early infancy or any long term measures of obesity (BMI, fat mass) or cardiovascular disease risk factors (such as elevated blood pressure).
Quality of the evidence
The interpretation of the review findings is limited by the existence in some of the trials of methodological weaknesses associated with potential for bias. The main concern is lack of evidence of use of methods to preserve allocation concealment in many of the trials. However, only one trial had substantial between-group differences in baseline demographics that is likely to be due to allocation bias (Jeon 2011). We elected to exclude one arm of this three-arm trial because of substantial differences in mean birth weight, gestational age, and proportion of growth-restricted infants. The other methodological limitation present in six of the trials was incomplete outcome assessment (loss to follow-up > 20%). In most of these trials, loss to follow-up was < 30% and was distributed evenly between intervention and control groups (Atkinson 1999; Koo 2006; Peng 2004; Picaud 2005). In two trials, loss to follow-up at 12 months assessment was > 50% (Agosti 2003; Carver 2001). However, these trial did not contribute substantially to any of the meta-analyses.
Potential biases in the review process
The main concern with the review process is the possibility that the findings are subject to publication and other reporting biases including more availability of numerical data for inclusion in meta-analyses in trials which reported statistically significant or clinically important effects (Hopewell 2009). We attempted to minimise this threat by searching the proceedings of the major international perinatal conferences to identify trial reports that are not (or not yet) published in full form in academic journals. However, we cannot be sure that other trials have been undertaken but not reported and the concern remains that such trials are less likely than published trials to have detected statistically significant or clinically important effects. The meta-analyses that we performed did not contain sufficient trials to explore symmetry of funnel plots as a means of identifying possible publication or reporting bias.
Implications for practice
The findings of this review do not support expert group and consensus recommendations that formula-fed preterm infants should receive a post-discharge formula for up to 12 months post-discharge (Dusick 2003; Kleinman 2004; Bhatia 2005; Carver 2005; Aggett 2006; Griffin 2007). In contrast, the available trial data indicate that feeding with "preterm formula", which is generally only licensed and available for in-hospital use, may increase weight, length and head circumference growth up to 12 to 18 months corrected age.
The infants who participated in the trials included in this review were fed ad libitum and the findings may not be applicable to infants who cannot feed ab libitum, for example because of oro-motor dysmotility or chronic lung disease.
Implications for research
Follow-up of infants who participated in the trials identified in this review might provide further data on the effect of this intervention on growth through later childhood, specifically whether final height is affected, on later neurodevelopmental outcomes, and on any long term effects on metabolic or cardiovascular outcomes (Euser 2005; Greer 2007). If further large randomised controlled trials to evaluate the effects of feeding preterm infants with nutrient-enriched formulae following hospital discharge are undertaken then it may be appropriate to include in any research efforts those preterm infants who are not able to feed ad libitum following hospital discharge, and who have extra metabolic demands, for example because of severe growth restriction or chronic lung disease. Trials should aim to assess long-term clinically important outcomes including final height and body composition and neurodevelopment (including cognitive and educational outcomes).
Further research is needed to determine which specific nutrients (including appropriate energy:protein balance) are key to promoting lean mass and linear growth and to improving developmental outcomes. As a first step, it may be worthwhile reviewing systematically trials that could not be included in this review because the nutrient-enriched formula differed only in protein and mineral content (but not energy) from standard term formula.
Editorial support of the Cochrane Neonatal Review Group has been funded with Federal funds from the Eunice Kennedy Shriver National Institute of Child Health and Human Development National Institutes of Health, Department of Health and Human Services, USA, under Contract No. HHSN275201100016C.
Data and analyses
- Top of page
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Sources of support
- Differences between protocol and review
- Index terms
Last assessed as up-to-date: 30 September 2011.
Protocol first published: Issue 2, 2004
Review first published: Issue 2, 2005
Contributions of authors
Lauren Young and Jessie Morgan undertook the electronic search and identified citations for possible inclusion. Lauren Young, Jessie Morgan and Felicia McCormick reviewed the citation list (title and abstract) for inclusion and undertook methodological appraisal, data extraction, entry and analysis. William McGuire acted as an arbiter for any disagreements, reviewed data entry and analysis and completed the review.
Declarations of interest
Sources of support
- Centre for Reviews and Dissemination, Department of Health Sciences, & Hull York Medical School, University of York, UK.
- Tenovus, Scotland, UK.
- NIHR, UK.Lauren Young and Jessie Morgan are NIHR Academic Clinical Fellows.
Differences between protocol and review
We elected to undertake separate comparisons of post-discharge formula and preterm formula versus standard term formula having previously specified a joint comparison with subgroup analysis.
Medical Subject Headings (MeSH)
Child Development [*physiology]; Dietary Proteins [administration & dosage]; Energy Intake [*physiology]; Infant Formula [*administration & dosage; standards]; Infant Nutritional Physiological Phenomena; Infant, Low Birth Weight [growth & development]; Infant, Newborn; Infant, Premature [*growth & development]; Patient Discharge; Randomized Controlled Trials as Topic
MeSH check words
* Indicates the major publication for the study