Timing of Breastfeeding Initiation and Exclusivity of Breastfeeding During the First Month of Life: Effects on Neonatal Mortality and Morbidity—A Systematic Review and Meta-analysis

Abstract

The purpose of this study was to review the evidence on the effect of initiation of breastfeeding early after birth and of exclusive breastfeeding during the first month in reducing neonatal mortality and morbidity. We searched Cochrane and PubMed databases for all available papers addressing our review questions and identified eleven papers. Data were extracted using a standard abstraction form. Evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation system. Meta-analysis was done using STATA 11.0. Early initiation of breastfeeding was associated with a reduced risk of neonatal mortality. Initiating breastfeeding after the first hour doubled the risk of neonatal mortality. Exclusively breastfed neonates had a lower risk of mortality and infection-related deaths in the first month than partially breastfed neonates. Exclusively breastfed neonates also had a significantly lower risk of sepsis, diarrhea and respiratory infections compared with those partially breastfed. The pooled evidence indicates that substantial benefits in reducing neonatal mortality and morbidity can be achieved with effective promotion of early initiation of breastfeeding and exclusive breastfeeding during the first month of life.

Appendices

Appendix 1: PubMed Search

Systematic Review PICO Questions

• 1. In all newborn infants (P), does initiation of breastfeeding within the first hour after birth (I) compared with later initiation of breastfeeding (C) reduce the risk of mortality and morbidity during the neonatal period (O)?

• 2. In all newborn infants (P), does exclusive breastfeeding during the first month of life (I) compared with predominant (C) or partial breastfeeding (C) reduce the risk of mortality and morbidity during the neonatal period (O)?

Search Criteria

(“breast feeding”[MeSH Terms] OR (“breast feeding”[MeSH Terms] OR (“breast”[All Fields] AND “feeding”[All Fields]) OR “breast feeding”[All Fields] OR “breastfeeding”[All Fields]) OR (“breast feeding”[MeSH Terms] OR (“breast”[All Fields] AND “feeding”[All Fields]) OR “breast feeding”[All Fields]) OR (“breast feeding”[MeSH Terms] OR (“breast”[All Fields] AND “feeding”[All Fields]) OR “breast feeding”[All Fields]) OR breastfed[All Fields] OR (“breast feeding”[MeSH Terms] OR (“breast”[All Fields] AND “feeding”[All Fields]) OR “breast feeding”[All Fields] OR (“breast”[All Fields] AND “fed”[All Fields]) OR “breast fed”[All Fields]) OR (“breast feeding”[MeSH Terms] OR (“breast”[All Fields] AND “feeding”[All Fields]) OR “breast feeding”[All Fields] OR (“breast”[All Fields] AND “fed”[All Fields]) OR “breast fed”[All Fields]) OR “colostrum”[MeSH Terms] OR (“colostrum”[MeSH Terms] OR “colostrum”[All Fields]) OR (“colostrum”[MeSH Terms] OR “colostrum”[All Fields] OR “colostrums”[All Fields])))

AND

((“infant, newborn”[MeSH Terms] OR (“infant”[MeSH Terms] OR “infant”[All Fields]) OR (“infant, newborn”[MeSH Terms] OR (“infant”[All Fields] AND “newborn”[All Fields]) OR “newborn infant”[All Fields] OR “newborn”[All Fields]) OR (“infant, newborn”[MeSH Terms] OR (“infant”[All Fields] AND “newborn”[All Fields]) OR “newborn infant”[All Fields] OR “newborns”[All Fields]) OR (“infant, newborn”[MeSH Terms] OR (“infant”[All Fields] AND “newborn”[All Fields]) OR “newborn infant”[All Fields] OR “neonate”[All Fields]) OR (“infant, newborn”[MeSH Terms] OR (“infant”[All Fields] AND “newborn”[All Fields]) OR “newborn infant”[All Fields] OR “neonates”[All Fields]) OR neonatal[All Fields] OR (“infant, newborn”[MeSH Terms] OR (“infant”[All Fields] AND “newborn”[All Fields]) OR “newborn infant”[All Fields] OR “baby”[All Fields] OR “infant”[MeSH Terms] OR “infant”[All Fields]) OR (“infant”[MeSH Terms] OR “infant”[All Fields] OR “babies”[All Fields]) OR “infant, newborn”[MeSH Terms])

AND

(“breast feeding”[MeSH Terms] OR (“breast feeding”[MeSH Terms] OR (“breast”[All Fields] AND “feeding”[All Fields]) OR “breast feeding”[All Fields] OR (“breast”[All Fields] AND “feeding”[All Fields] AND “exclusive”[All Fields])) OR “breast feeding”[MeSH Terms] OR (“breast feeding”[MeSH Terms] OR (“breast”[All Fields] AND “feeding”[All Fields]) OR “breast feeding”[All Fields] OR (“breastfeeding”[All Fields] AND “exclusive”[All Fields])) OR “breast feeding”[MeSH Terms] OR (“breast feeding”[MeSH Terms] OR (“breast”[All Fields] AND “feeding”[All Fields]) OR “breast feeding”[All Fields] OR (“exclusive”[All Fields] AND “breast”[All Fields] AND “feeding”[All Fields]) OR “exclusive breast feeding”[All Fields]) OR “breast feeding”[MeSH Terms] OR (“breast feeding”[MeSH Terms] OR (“breast”[All Fields] AND “feeding”[All Fields]) OR “breast feeding”[All Fields] OR (“exclusive”[All Fields] AND “breastfeeding”[All Fields]) OR “exclusive breastfeeding”[All Fields]) OR initiate[All Fields] OR initiation[All Fields] OR begin[All Fields] OR beginning[All Fields] OR start[All Fields] OR starting[All Fields] OR started[All Fields] OR initiated[All Fields] OR begun[All Fields] OR initiating[All Fields])

NOT “addresses”[Publication Type] NOT “autobiography”[Publication Type] NOT “bibliography”[Publication Type] NOT “biography”[Publication Type] NOT “case reports”[Publication Type] NOT “clinical conference”[Publication Type] NOT “collected works”[Publication Type] NOT “comment”[Publication Type] NOT “congresses”[Publication Type] NOT “consensus development conference”[Publication Type] NOT “consensus development conference, nih”[Publication Type] NOT “dictionary”[Publication Type] NOT “directory”[Publication Type] NOT “editorial”[Publication Type] NOT “electronic supplementary materials”[Publication Type] NOT “ephemera”[Publication Type] NOT “festschrift”[Publication Type] NOT “guideline”[Publication Type] NOT “historical article”[Publication Type] NOT “in vitro”[Publication Type] NOT “interactive tutorial”[Publication Type] NOT “interview”[Publication Type] NOT “lectures”[Publication Type] NOT “legal cases”[Publication Type] NOT “legislation”[Publication Type] NOT “letter”[Publication Type] NOT “news”[Publication Type] NOT “newspaper article”[Publication Type] NOT “practice guideline”[Publication Type] NOT “review”[Publication Type]

Limits applied: humans, new born: birth-1 month.

Characteristics of the Included Studies

Arifeen 2001

Methods	Cohort study
Participants	1,677 singleton births included into the study born to a cohort of mothers resident in 5 sub-districts of Dhaka city, Bangladesh between November 1993 and June 1995 Loss to follow up: <12 %
Intervention	Type of breastfeeding: [1] exclusively breastfed, [2] predominantly breastfed and [3] partial or not breastfed
Outcome	Neonatal mortality was attributed to all causes, ARI and diarrhea

Risk of bias

Items	Author’s judgment	Description
Confounding	Differences in height and education	Controlled

Cushing 1997

Methods	Cohort study
Participants	1,202 babies born between January 1, 1988 and June 30, 1990 in Albuquerque, New Mexico, United States Inclusion criteria: Mother’s age above 18 years, English speaking non-smoker and with access to telephone at home Exclusion criteria: If parents plan to put the infants in the full time day care and if the family plans to move from Albuquerque within the next 18 months Loss to follow up: 13 %
Intervention	Type of breastfeeding: [1] fully breastfed, [3] partially breastfed and [3] not breastfed
Outcome	Lower, upper and all respiratory illness

Risk of Bias

Item	Author’s judgment	Description
Confounding	Differences in income and maternal education	Controlled
Misclassification	Mild	Possible due to recall (telephone interviews) and errors in mother’s filling of the diaries

Clavano 1982

Methods	Cohort study
Participants	9,886 charts reviewed from babies delivered between January 1973 and April 1977 in a hospital in the Philippines Inclusion criteria: Newborns Exclusion criteria: No feeding record available, newborns deaths due to different causes like those acutely ill at birth, immature and having congenital abnormalities Loss to follow up: <3 %
Intervention	Type of breastfeeding: [1] breastfed, [2] mixed fed and [3] formula fed
Outcomes	Neonatal mortality and morbidity attributed to diarrhea, sepsis and oral thrush

Risk of bias

Item	Author’s judgment	Description
Confounding	Groups reported as comparable	Data was not provided

Garcia 2011

Methods	Cohort study as part of randomized, placebo controlled trial for Vitamin A
Participants	10,464 newborns in a rural village in Tamil Nadu, India Inclusion criteria: Newborns between 2 and 28 days old and breastfeeding initiation time could be determined Exclusion criteria: Women delivered 20 km or more outside the study area, first visit >7 days after birth Loss to follow up: <1 %
Intervention	Breastfeeding initiation time: [1] <12 h after birth, [2] 12–24 h after birth and [3] >24 h after birth
Outcome	Neonatal mortality

Risk of bias

Item	Author’s judgment	Description
Reverse causation bias	Mild	Risk reduced as deaths in the first 48 h excluded
Recall bias	Low	Only infants with visits done in the first 7 days after birth included
Confounding	Birth weight, sex, care at birth, vitamin A status, household head’s occupation	Controlled

Edmond 2006

Methods	Cohort study as part of a large Vitamin A trial
Participants	10,947 newborns in rural Ghana Inclusion criteria: Infants surviving to day 2 and who were breastfed successfully Exclusion criteria: Multiple births, non-initiators and those interviewed outside the neonatal period Loss to follow up: 4.4 %
Intervention	Initiation of breastfeeding: [1] Early versus [2] late Type of breastfeeding: [1] Exclusive, [2] partial and [3] predominant
Outcome	Neonatal mortality

Risk of bias

Item	Author’s judgment	Description
Reverse causality	Mild	Risk was reduced as all deaths in the 1st 2 days were excluded
Misclassification	Moderate	24 h recall for feeding status at 1 month, but 4 weeks recall for timing of initiation
Confounding	Gender, birth size, gestational age, congenital anomaly, health on day of birth, health at time of interview, mother’s health at delivery, mother’s age, parity, education, cash income, household water supply, place of defecation, antenatal visits, place of birth and presence of birth attendant	Controlled

Edmond 2007

Methods	Cohort study as part of a large Vitamin A trial
Participants	10,942 newborns in rural Ghana Inclusion criteria: Infants surviving to day 2 and who were breastfed successfully Exclusion criteria: Multiple births, non-initiators and those interviewed outside the neonatal period Loss to follow up: <5 %
Intervention	Initiation of breastfeeding: [1] Early versus [2] late Type of breastfeeding: [1] Exclusive, [2] partial and [3] predominant
Outcome	All infectious disease related neonatal mortality

Risk of bias

Item	Author’s judgment	Description
Reverse causality	Mild	Risk was reduced as all deaths in 1st 2 days were excluded
Misclassification	Moderate	24 h recall for feeding status at 1 month, but 4 weeks for timing of initiation
Confounding	Gender, birth size, gestational age, congenital anomaly, health on day of birth, health at time of interview, mother’s health at delivery, mother’s age, parity, education, cash income, household water supply, place of defecation, antenatal visits, place of birth and presence of birth attendant	Controlled

Kasla 1995

Methods	Cohort study
Participants	537 newborns in a hospital in Bombay, India Inclusion criteria: Newborns Exclusion criteria: Birth weight <2,000 g, seriously ill in early neonatal period Loss to follow up: 12 %
Intervention	Type of breastfeeding: [1] Exclusive, [2] artificial and [3] mixed
Outcomes	Total morbidity and morbidity due to gastrointestinal and respiratory infections

Risk of bias

Item	Author’s judgment	Description
Misclassification	Mild	Disease classification based on mother’s recall and classification
Confounding	Possible confounding not described	Not adjusted in the analysis

Lopez-Alarcon 1997

Methods	Cohort study
Participants	216 newborns in a community and a hospital in Mexico City, Mexico Inclusion criteria: Singleton, full-term, healthy and weighing >2,500 g at birth Exclusion criteria: Congenital malformations and those who moved out Loss to follow up: 22 %
Intervention	Type of breastfeeding: [1] Exclusively breastfed, [2] partially breastfed and [3] formula fed
Outcomes	Diarrhea and acute respiratory infection related morbidity

Risk of bias

Item	Author’s judgment	Description
Misclassification	Mild	Babies shifted during the 6 month period in the three feeding groups Misclassification of outcome unlikely as the diagnosis was clear and the physicians were blinded to the feeding history
Confounding	Possible—differences by feeding groups not described	Not controlled

Mullany 2008

Methods	Cohort study
Participants	23,662 newborns between 2002 and 2006 as part of a randomized control trial in a rural community in Southern Nepal Inclusion criteria: Newborns Exclusion criteria: If two or more of the following conditions were present 1. Difficulty breathing 2. Stiffness of the back or convulsions 3. Dysentery 4. Presence of watery stools five or more times within 24 h 5. Severe chest in drawing 6. Axillary temperature >37.8 °C 7. Respiratory rate <70 breath/min Loss to follow up: 1.4 %
Interventions	Initiation on breastfeeding: [1] Early versus [2] late
Outcomes	Neonatal mortality and morbidity

Risk of bias

Item	Author’s judgment	Description
Reverse causation bias	Low	Early deaths excluded
Recall bias	Low	Feeding classification based on 24 h recall
Confounding	Birth weight, gestation, sex, maternal literacy, chlorhexidine use, maternal hand washing, death of siblings, parity, maternal fever, ethnicity	Controlled

Sinha 2003

Methods	Case control study
Participants	933 infants born between 1990 and 1998 in a hospital in Boston, Massachusetts, United States Inclusion criteria: Infants delivered ≥37 weeks gestation with respiratory tract infection (LRTI but not pneumonia) diagnosed after discharge Exclusion criteria: Pneumonia and delivery <37 weeks Loss to follow up: 1 %
Intervention	Type of breastfeeding: [1] Exclusive breastfed, [2] mixed breastfed and [3] not breastfed
Outcomes	Neonatal morbidity

Risk of bias

Item	Author’s judgment	Description
Selection bias or measurement bias	Mild	Electronic search of feeding records
Confounding and adjustment	Year of birth, season of birth, presence of siblings, and socioeconomic status	Controlled

Victora 1987

Methods	Case control study
Participants	510 Neonates in urban area in Southern Brazil Inclusion criteria: Infant deaths 7–364 days of life Exclusion criteria: Very low birth weight (<1,500 g), multiple births, more than 15 days in hospital immediately before death and congenital malformation or cerebral palsy Loss to follow up: <2 %
Intervention	Cases were deaths due to diarrhea, respiratory and other infections (meningitis, skin infections, measles, whooping cough, neonatal sepsis and tuberculosis Controls were neighbors aged 7–364 days Type of feeding: [1] Breastfeeding, [2] mixed and [3] artificial
Outcomes	Neonatal deaths

Risk of bias

Item	Author’s judgment	Description
Interviewer bias	Mild	Reduced as information on outcome (diarrhea) was collected at the end of interview
Reverse causality	Mild	Feeding classified as status before illness
Confounding	Age, social status, birth weight, type of housing, piped water, birth interval, maternal education, family income	Controlled