Cessation of exclusive breastfeeding (EBF) with early introduction of complementary food provides additional calories for catch-up growth but may also increase the risk of adverse outcomes. The objective of this study was to assess effects of exclusive breastfeeding for less than 6 months compared with 6 months in preterm and low birth weight infants.
Data sources include Medline, Scopus, Web of Science, CINAHL, and Index Medicus through June 30, 2021. Study selection includes randomized trials and observational studies. Primary outcomes were mortality, morbidity, growth, and neurodevelopment. Data were extracted and pooled using random-effects models. The Cochrane Risk of Bias 2 tool was used to assess the risk of bias of included studies.
A total of 2 studies of 307 preterm or low birth weight infants were included. None of the study results could be pooled. Both studies compared EBF for 4 months to 6 months. Growth was similar between the 4-month and 6-month EBF groups for the following outcomes: weight-for-age z-score at corrected age 12 months (mean [standard deviation], 4-month group: −1.7 [1.1], 6-month group: −1.8 [1.2], 1 study, 188 participants, low certainty evidence), absolute weight gain (gram) from 16 to 26 weeks of age (4-month group: 1004 [366], 6-month group: 1017 [350], 1 study, 119 participants, very low certainty evidence), and linear growth gain (cm) from 16 to 26 weeks of age (4-month group: 4.3 [0.9], 6-month group: 4.5 [1.2], 1 study, 119 participants, very low certainty evidence). There were no apparent differences in reported morbidity symptoms. No difference in the timing to achieve motor development milestones between the 2 groups was found (1 study; 119 participants, very low certainty evidence). A limited number of studies prevented data pooling.
The evidence is very uncertain about the effect of exclusive breastfeeding for less than 6 months for preterm and low birth weight infants. Further studies are warranted to better answer this question.
The World Health Organization (WHO) and the United Nations Children’s Fund, along with many national and pediatric professional societies, recommend exclusively breastfeeding for 6 months, followed by the introduction of complementary feeding for term infants.1,2 The WHO also recommends that low birth weight (LBW) infants be exclusively breastfed for 6 months, though this does not apply to sick infants or those weighing less than 1000 g, and there is no global guidance regarding the optimal duration of exclusive breastfeeding and the ideal timing to introduce complementary feeding for preterm infants.2,3 A 1994 United Kingdom working group recommended preterm infants start complementary feeding when reaching at least 5 kg of weight, having lost the extrusion reflex, and are able to eat from a spoon.4 A review by Palmer et al suggested initiating solid food at 13 weeks of corrected age.5 Existing evidence on when to initiate complementary feeding for preterm infants is inconclusive.6,7
The optimal duration of exclusive breastfeeding for preterm and LBW infants remains controversial because of the potential for both benefits and harms. Early complementary feeding introduces additional calories, which may promote catch-up growth for preterm infants as milk feeds alone may provide insufficient nutrition. One randomized clinical trial found that preterm infants who were randomized to initiate complementary feeding at 13 weeks postnatal age had higher growth velocity by the corrected age of 12 months compared with those randomized to weaning at 17 weeks postnatal age.8 However, some observational studies suggest the association between early introduction of complementary feeding in infants and obesity later in childhood.9,10 Although various studies have linked breastfeeding with lower risks of allergic diseases, a 2012 meta-analysis including 3 studies found no additional benefit of exclusively breastfeeding for more than 3 or 4 months in reducing the risk of allergies, asthma, or eczema.11 A Cochrane review found no link between breastfeeding for 6 months and reduced risks of allergic diseases in children at aged 5 to 7.12 Further, a landmark trial in the United Kingdom that assessed introduction of allergenic foods to infants at 3 months versus 6 months of age found that earlier introduction of peanuts and eggs may reduce the risk of allergies to these foods by the age of 3.13 Importantly, exclusive breastfeeding for at least 6 months clearly reduces the risk of gastrointestinal infections and likely the risk of respiratory infections.12 Recommendations must consider both the potential risks and benefits in the short and long run.
Previous studies have focused on the timing of introducing complementary feeding for preterm or LBW infants in general, rather than focusing on those who were exclusively breastfed.8,14,15 As the current global guidelines recommend exclusive breastfeeding for 6 months, the critical question remains: what is the optimal duration of exclusive breastfeeding for preterm or LBW infants? The objective of this study was to evaluate the effect of exclusive breastfeeding for less than 6 months, versus 6 months, on critical outcomes, including mortality, morbidity, growth, and neurodevelopment outcomes among preterm and LBW infants.
Methods
The protocol of this review was registered in the International Prospective Register of Systematic Reviews (PROSPERO) database (Registration number CRD42021240480). This review was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.16
Inclusion and Exclusion Criteria
We included randomized control trials, nonrandomized trials, observational, and mixed-method studies that examined the effects of exclusive breastfeeding for less than 6 months, versus for 6 months, on the outcomes of interest in preterm infants (delivered at gestational age less than 37 weeks) or LBW infants (birth weight less than 2500 g). Only peer-reviewed articles with full text were included. Papers in all languages were eligible.
Outcome
Primary outcomes were mortality, morbidity, growth, and neurodevelopment outcomes assessed at 6, 12, or 24 months of chronological age or corrected age and the latest follow-up when data are available (Appendix 1). Morbidities of interest included infections, such as diarrhea, acute respiratory tract infection, acute lower respiratory tract infection, influenza, pneumonia, and sepsis, as well as hospitalization history and allergic diseases such as eczema. Growth outcomes included weight-for-age z-score (WAZ), length-for-age z-score, weight-for-length/height z-score, absolute weight gain (gram), absolute linear growth gain (cm), weight gain rate (g/kg per day), and linear growth rate (cm per week and cm per month). Neurodevelopment outcomes were assessed based on any standardized or validated test for cognitive function, motor, language, and emotional outcomes and presented by standardized score (z score).
Search Strategy
The full search strategy is outlined in Appendix 2. We searched the following databases: Medline, Scopus, Web of Science, CINAHL, and Index Medicus and conducted manual reference checks of existing reviews.
Study Selection and Data Extraction
Three reviewers (M.L., A.F., W.C.Y.) independently screened titles and abstracts based on the predetermined inclusion and exclusion criteria for articles identified from the initial search. Duplicates, irrelevant interventions, and wrong study designs were removed. Two reviewers (M.L., A.F.) performed full text screening to assess the eligibility of studies by a checklist and used web-based software, Covidence, to manage the screening process.17 Discrepancies at both stages were resolved by a third reviewer (E.R.S.). Three reviewers (M.L., A.F., W.C.Y.) performed data extraction independently using an online data extraction sheet and cross-checked the extracted data to ensure accuracy. Detailed study information including authors, publication year, location, number of participants, and gestational age and birth body weight of participants were collected.
Assessment of Risk of Bias
Two reviewers (W.C.Y., M.L.) assessed the risk of bias of trials using the Cochrane Risk of Bias 2 tool.18 A third reviewer (E.R.S.) resolved disagreements between reviewers. We also adopted the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach to assess the overall quality of the evidence using The GRADEpro GDT software.19,20
Statistical Analysis
For outcomes with at least 2 studies, we pooled data using conventional random-effects meta-analysis. For binary outcomes, risk ratios (RR) with 95% confidence intervals (CI) were used; for continuous outcomes, we intended to pool the mean difference between intervention groups. Subgroup analysis by birth weight (eg, <1000 g, 1000 to 1499 g, 1500 to 2500 g) and gestational age (eg, <28 weeks, 28 to <32 weeks, 32 to 36 weeks, >37 weeks) would be performed to assess heterogeneity when data were available.
Results
Included Studies and Characteristics
We included a total of 2 studies with 307 preterm or LBW infants21,22 (Appendix 3, PRISMA flow diagram). Characteristics of the 2 trials are shown in Appendix 4. The study by Gupta et al, including preterm infants less than 34 weeks gestational age, randomized infants to the introduction of complementary food at corrected age 4 and 6 months.21 As the study included infants who were exclusively, predominantly, and partially breastfeeding at the time of randomization, the only outcome data that was available for exclusively breastfeeding preterm infants was WAZ. Dewey et al included term infants with birth weight between 1500 to 2500 g.22 Dewey and colleagues published 3 relevant papers based on the data from the same trial, which evaluated the effect of introducing complementary feeding at 4 and 6 months of age among exclusively breastfeeding low birth weight infants.22–24 The Dewey 2004 study was not included as it did not examine the outcome of interest determined by our protocol.24
Risk of Bias
Outcomes
Table 1 summarizes our study results. The data could not be pooled as outcomes were measured and reported differently. Regarding the quality of evidence, all outcomes were downgraded for inconsistency because only 1 study was available and for imprecision because of wide confidence intervals crossing the line of no effect. Hence, the overall quality of the evidence was low to very low. Between exclusive breastfeeding for 4 months and for 6 months, the mean difference of WAZ at corrected age 12 months was 0.1 (95% CI −0.2 to 0.4, I2 = NA, low certainty evidence, 1 trial, 188 participants). The mean differences of weight gain and length gain from 16 to 26 weeks of chronological age were −13 g (95% CI −143 to 117, I2 = NA, very low certainty evidence, 1 trial, 119 participants) and −0.2 cm (95% CI −0.6 to 0.2, I2 = NA, very low certainty evidence, 1 trial, 119 participants).
Summary of Findings for Critical Outcomes: Exclusive Breast Feeding for <6 Months Compared With Exclusive Breast Feeding for 6 mo for Preterm and LBW Infants
| Outcomes . | N of Participants (Studies)Follow-up . | Certainty of the Evidence (GRADE) . | Relative Effect (95% CI) . | Anticipated Absolute Effects . | |
|---|---|---|---|---|---|
| Risk With Exclusive Breast Feeding for 6 mo . | Risk Difference With Exclusive Breast Feeding for <6 mo . | ||||
| Mortality | No trials | ||||
| Morbidity: % days with diarrhea at 26 wk chronological age | 119 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean % days with diarrhea 5.4% | 2.6% days fewer (5.2 fewer to 0) |
| Morbidity: % days with fever at 26 wk chronological age | 119 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean % days with diarrhea 8.0% | 0.7% days fewer (3.4 fewer to 2 more) |
| Wt for age z score at corrected age 12 mo | 188 (1 RCT) | ⨁⨁◯◯ Lowa,b | — | Mean wt for age z score at CA 12 mo was −1.8 WAZ | 0.1 WAZ higher (0.2 lower to 0.4 higher) |
| Wt gain in grams from 16 to 26 wk of chronological age | 119 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean wt gain in grams from 16 to 26 wk of age was 1017 gm | 13 gm lower (143 lower to 117 higher) |
| Length gain in cm from 16 to 26 wk of age | 119 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean length gain in cms from 16 to 26 wk of age was 4.5 cm | 0.2 cm lower (0.6 lower to 0.2 higher) |
| Motor development milestone: age in months when reported to be able to raise head | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: raise head was 1.0 month of age | 0 mo of age (0.3 fewer to 0.3 higher) |
| Motor development milestone: age in months when reported to be able to raise head and chest | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: raise head and chest was 1.9 month of age | 0.1 mo of age fewer (0.7 fewer to 0.5 more) |
| Motor development milestone: age in months when reported to be able to roll over | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: roll over was 3.8 month of age | 0 mo of age (0.7 fewer to 0.7 higher) |
| Motor development milestone: age in months when reported to be able to crawl | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: crawl was 6.8 month of age | 0.6 mo of age more (0.1 fewer to 1.3 more) |
| Motor development milestone: age in months when reported to be able to sit from lying position | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: sit from lying position was 7.4 month of age | 0.6 mo of age more (0 to 1.2 more) |
| Infant can walk by the age of 12 mo | 99 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | RR 1.47 (0.69 to 3.13) | 180 per 1000 | 85 more per 1000 (56 fewer to 383 more) |
| Outcomes . | N of Participants (Studies)Follow-up . | Certainty of the Evidence (GRADE) . | Relative Effect (95% CI) . | Anticipated Absolute Effects . | |
|---|---|---|---|---|---|
| Risk With Exclusive Breast Feeding for 6 mo . | Risk Difference With Exclusive Breast Feeding for <6 mo . | ||||
| Mortality | No trials | ||||
| Morbidity: % days with diarrhea at 26 wk chronological age | 119 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean % days with diarrhea 5.4% | 2.6% days fewer (5.2 fewer to 0) |
| Morbidity: % days with fever at 26 wk chronological age | 119 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean % days with diarrhea 8.0% | 0.7% days fewer (3.4 fewer to 2 more) |
| Wt for age z score at corrected age 12 mo | 188 (1 RCT) | ⨁⨁◯◯ Lowa,b | — | Mean wt for age z score at CA 12 mo was −1.8 WAZ | 0.1 WAZ higher (0.2 lower to 0.4 higher) |
| Wt gain in grams from 16 to 26 wk of chronological age | 119 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean wt gain in grams from 16 to 26 wk of age was 1017 gm | 13 gm lower (143 lower to 117 higher) |
| Length gain in cm from 16 to 26 wk of age | 119 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean length gain in cms from 16 to 26 wk of age was 4.5 cm | 0.2 cm lower (0.6 lower to 0.2 higher) |
| Motor development milestone: age in months when reported to be able to raise head | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: raise head was 1.0 month of age | 0 mo of age (0.3 fewer to 0.3 higher) |
| Motor development milestone: age in months when reported to be able to raise head and chest | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: raise head and chest was 1.9 month of age | 0.1 mo of age fewer (0.7 fewer to 0.5 more) |
| Motor development milestone: age in months when reported to be able to roll over | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: roll over was 3.8 month of age | 0 mo of age (0.7 fewer to 0.7 higher) |
| Motor development milestone: age in months when reported to be able to crawl | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: crawl was 6.8 month of age | 0.6 mo of age more (0.1 fewer to 1.3 more) |
| Motor development milestone: age in months when reported to be able to sit from lying position | 108 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | — | Mean motor development milestone: sit from lying position was 7.4 month of age | 0.6 mo of age more (0 to 1.2 more) |
| Infant can walk by the age of 12 mo | 99 (1 RCT) | ⨁◯◯◯ Very lowa,b,c | RR 1.47 (0.69 to 3.13) | 180 per 1000 | 85 more per 1000 (56 fewer to 383 more) |
Patient or population: preterm and low birth wt infants. Setting: any. Intervention: exclusive breast feeding for <6 mo. Comparison: exclusive breast feeding for 6 mo. The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). GRADE Working Group grades of evidence: high certainty, we are very confident that the true effect lies close to that of the estimate of the effect; moderate certainty, we are moderately confident in the effect estimate and the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different; low certainty, our confidence in the effect estimate is limited and the true effect may be substantially different from the estimate of the effect; and very low certainty, we have very little confidence in the effect estimate and the true effect is likely to be substantially different from the estimate of effect. CI, Confidence interval; RR, Risk ratio; —, not applicable.
Only one study is available, so inconsistency was downgraded by one level because it could not be evaluated.
Wide confidence intervals crossing line of no effect, imprecision downgraded 1 level.
Lack of allocation concealment as randomization was performed by week of birth in the Dewey 1999 study.
The mean difference in the percentage of days with diarrhea from 16 to 26 weeks of chronological age was −2.6% of days (95% CI −5.2 to 0, I2 = NA, very low certainty evidence, 1 trial, 119 participants) and that with fever was −0.7% of days (95% CI −3.4 to 2.0, I2 = NA, very low certainty evidence, 1 trial, 119 participants). Regarding motor development milestones, the relative risk of parents reporting their infants to be able to walk by 12 months of age was 1.47 (95% CI 0.69 to 3.13, I2 = NA, very low certainty evidence, 1 trial, 108 participants). The evidence was very uncertain as to whether there was, on average, any difference between groups in the age in months when the infants were reported to be able to raise their head and chest, roll over, crawl, and sit from lying position (Table 1, all very low certainty evidence, 1 trial, 108 participants). There were no data on mortality, other growth outcomes such as head circumference, or standardized neurodevelopmental outcomes.
Discussion
Our systemic review of 2 studies and 307 infants found that the evidence was very uncertain regarding whether there were any differences in mortality, morbidity, growth, and neurodevelopment outcomes between those who were exclusively breastfed for less than 6 months and for 6 months. Based on low certainty evidence, exclusive breastfeeding for less than 6 months may have little to no difference on WAZ in preterm or LBW infants.
When considering whether to recommend exclusive breastfeeding for less than 6 months for preterm and LBW infants, it is important to balance potential benefits with unintended consequences such as increased risks of mortality and infectious disease in the short-term or adverse cardiometabolic outcomes and allergies later in life. Studies linking early introduction of complementary food and childhood obesity were mainly restricted to term infants; further research specific to preterm and LBW infants is needed.9,10 Among term infants, there is clear evidence of increased risks of mortality, diarrhea, and respiratory infections associated with formula or solid food feeding as compared with breastfeeding, and this data informs global guidance in favor of exclusive breastfeeding (EBF) for 6 months.25–27 Although we found very uncertain evidence about the optimal duration of EBF for preterm infants to prevent infection, the randomized trial by Gupta, including both EBF and non-EBF preterm infants, showed that infants randomized to the earlier introduction of complementary feeding at 4 months were more likely to be hospitalized, in many cases because of infectious disease.21 We could not include this data in our meta-analysis because the data were not stratified by the EBF status.
Anemia and micronutrient status are also important outcomes related to the duration of EBF among preterm infants, who are especially susceptible to anemia. The trial by Dewey et al included in our review found that among LBW, term infants who were not provided iron supplements, the 6-month EBF group was more likely to be iron deficient (6-month EBF group vs 4-month EBF group: 54.8% versus 15.4%) and anemic (6-month EBF group vs 4-month EBF group: 21.3% versus 2.2%). There were no differences in vitamin A B12, folate, or zinc status between the 2 groups.24 An earlier trial in Honduras among both LBW and non-LBW infants similarly found that infants in the 6-month EBF group had lower mean hemoglobin and ferritin concentrations at 6 months and were more likely to be anemic (RR 1.20; 95% CI 0.91 to 1.58) and iron-deficient (RR 2.93; 95% CI 1.13 to 7.56).28 A 2012 Cochrane review, including the Honduras data, concludes that EBF for 6 months without iron supplementation may compromise infant hematologic status in term infants in contexts where newborn iron stores are low.12 Additional evidence specific to preterm infants is needed.
Importantly, the timing of the introduction of complementary feeding is not an exactly overlapping question with the optimal duration of EBF. Early introduction of complementary foods can mean early cessation of breast milk, formula, or both. There are several studies evaluating the timing of introducing complementary food for preterm infants that could not be included in our review because their study populations included both exclusively breastfed and nonexclusively breastfed infants or because neither group was exclusively breastfed for 6 months. Morgan et al evaluated the effect of weaning before or after 12 weeks of age on growth at 18 months among preterm infants and found no difference.14 Marriott et al compared the growth velocity between preterm infants weaned at 13 and 17 weeks of postnatal age, but the analysis by milk feeds was unavailable.8 To better address the research and policy question at hand, future research could use existing cohort data to examine the dose-response effect of earlier breastfeeding cessation by the number of weeks or months infants are exclusively breastfed compared with infants who are exclusively breastfed for 6 months.
Our study has some limitations. First, there were only 2 eligible studies and none of the data could be pooled. Second, by strictly adhering to our protocol and requiring 6 months of EBF as the comparison group, we excluded studies that might be potentially informative about the optimal duration of EBF in preterm and LBW infants. Third, we did not include anemia and iron status as an outcome of interest in our review, and thus we can’t be sure if there is additional evidence regarding hematologic data available for preterm and LBW infants. However, a posthoc search did not reveal any additional evidence.
Conclusions
Based on very uncertain evidence, we found that preterm and LBW infants who are exclusively breastfed for less than 6 months may experience little to no difference in growth outcomes compared with infants exclusively breastfed for 6 months. Evidence regarding other outcomes is unavailable. As the evidence is extremely limited, further studies, including reanalysis of existing data, are warranted to better answer this question.
Drs Yang and Smith conceptualized and designed the study, conducted the initial analyses, and drafted the initial manuscript; Ms Fogel, Ms Lauria, and Ms Ferguson designed the data collection instruments, collected data, conducted the initial analyses; and all authors reviewed and revised the manuscript, approved the final manuscript as submitted, and agreed to be accountable for all aspects of the work.
This study is registered at Prospero https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=240480 (identifier, CRD4202124048).
FUNDING: All phases of this study were supported by the World Health Organization (WHO) Department of Maternal, Newborn, Child, Adolescent Health and Aging. The sponsor (WHO) commissioned the review for a guideline development group meeting for development of WHO recommendations on care of the preterm or low birth weight infant. The sponsor assisted in formulating the research question and provided inputs on the synthesis of the results and manuscript.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest to disclose.
