Prospective Cohort Study

B - Click here for explanation of classification scheme.

POSITIVE: See Quality Criteria Checklist below.

The aim of this prospective observational study was to elucidate the feeding routines at a maternity unit and to analyze the risk of not being breast fed at three months of age in relation to feeding patterns in the early neonatal period, while simultaneously taking into account a number of characteristics of the mother and child.

During the period from February to April 1990, 583 children were born at the University Hospital, Umeå, Sweden. 521 children (243 boys and 278 girls) born at the hospital were included in the study. One-quarter of the children received supplementary feeds on the third day of life, the indications for this being birth weight less than 3.0kg, maternal diabetes or gestational diabetes, “insufficient amounts of milk” or fussiness.

53 children who were treated entirely in the neonatal ward and the nine others who were discharged to other hospitals were excluded from the study.

Design

Prospective Cohort Study

 

Statistical Analysis

• The relative risks of not being breast-fed at three months of age were estimated by calculation of odds ratios (OR) and confidence limits (CL)
• When simultaneous adjustments were made for several potential confounders, multiple logistic regression analysis was used using EGRET software program.

Timing of Measurements

• Information regarding pregnancy, delivery and early neonatal feeding was registered by the maternity unit staff and cross-checked by a pediatrician when the mother and baby were discharged from the hospital
• A questionnaire regarding feeding patterns at one, two, three and four months was prospectively completed by the parents
• Missing information was obtained by telephone interviews
• One-quarter of the children received supplementary feeds on the third day of life, the indications for this being birth weight less than 3.0kg, maternal diabetes or gestational diabetes, “insufficient amounts of milk” or fussiness.

Dependent Variables

 

• Exclusively breast-fed: Infant received only breast milk from its mother, or expressed breast milk, and no other liquids or solids, with the exception of drops or syrup containing vitamins, mineral supplements or medicines
• Breast-fed: Infant received breast milk, but was not allowed to receive any food or liquid including non-human milk
• Supplementary donor’s milk: Infant received human milk from the pooled milk bank.

 

Independent Variables

 

• Mother’s age (years)
• Mother smoking
• Earlier births
• Delivery
• Gestational age (weeks)
• Birth weight (g)
• Sex
• Admitted neonatal ward initially
• Supplementary feeding
• Used breast  pump
• Used oxytocin spray
• Weight loss >=10%.

Control Variables

 All independent variables were in multiple logistic regression model.

 

Initial N

583

Attrition (final N)

521 children (243 boys and 278 girls). Note: Feeding information after the neonatal period were missing in 6%.

Age

See table 2

Ethnicity

Not described

Other relevant demographics

Location

Umea, Sweden

 

The indications for pediatricians to prescribe supplementary feeding are listed in Table 1. During the first day, maternal diabetes or gestational diabetes were major reasons, while a pronounced initial weight loss dominated after a few days. Man children received supplementary donor’s milk or formula for no specific reasons, sometimes indicated as “insufficient amounts of milk”

Table 1: Proportion of children receiving supplementary donor’s milk or formula in relation to different indications (percentage of total number of children for each day).

Day
Indication	1	2	3	4	5	6
Birth weight<3kg	2.9	4.7	5.5	3.2	1.9	0.0
Birth weight>4.5kg	0.2	0.4	0.4	0.4	0.2	0.9
Initial weight loss	0.0	0.2	1.6	4.4	2.6	0.9
Maternal diabetes or gestational diabetes	5.7	6.0	5.5	3.6	1.4	0.9
Jitteriness	0.2	1.0	1.2	0.4	0.0	0.0
Fussiness	0.4	1.4	2.3	1.2	0.5	0.9
Insufficient amount of milk	2.0	2.1	3.7	3.4	5.6	8.4
Other, unknown	1.2	1.8	2.2	2.4	2.8	6.7
Percent of children receiving supplementary feeding	12.5	17.5	22.3	18.6	15.0	18.7
Total number of children in the maternity unita	511	513	511	505	426	225

 ^a 35 children were partly treated in the neonatal unit, usually the first one to two days.

In the bi-variate analysis, an increased risk of not being breast-fed at three months was related to young maternal age, smoking, Cesarean section, supplementation with door’s milk or formula at the maternity unit, the use of oxytocin spray and an initial weight loss ≥10% of birth weight.

Table 2: Relative risk of not being breast fed at three months in relation to early feeding characteristics of the mother, the delivery and the child.

Characteristics		Non-breast fed	Breast-fed	OR	95% CL
Mother’s age (years)	25-	72	336	1.0
	<25	25	57	2.05	1.21-3.47
Mother’s Smoking	no	61	309	1.0
	yes	35	56	3.17	1.94-5.16
Earlier births	Yes	60	253	1.0
	No	37	139	1.12	0.71-1.78
Delivery	Vaginal	82	366	1.0
	Caesarean	14	22	2.84	1.43-5.65
Gestational age (weeks)	38-	83	347	1.0
	<38	11	28	1.64	0.79-3.42
Birth weight (g)	<3,000	13	48	1.0
	3,000-4,499	83	337	0.91	0.47-1.76
	4,500-	2	9	0.82	0.16-4.31
Sex	Boy	49	184	1.0
	Girl	49	210	0.88	0.56-1.36
Admitted neonatal ward initially	No	88	372	1.0
	Yes	10	22	1.92	0.89-4.16
Supplementary feeding	No	44	292	1.0
	Yes	54	102	3.51	2.26-5.47
Used breast pump	No	64	284	1.0
	Yes	33	108	1.36	0.84-2.18
Used oxytocin spray	No	74	348	1.0
	Yes	24	46	2.45	1.43-4.22
Weight loss>=10%	No	70	345	1.0
	Yes	23	40	2.83	1.62-4.95

 

Many of these factors in Table 2 are associated with each other. In order to control for confounding, multiple logistic regression including all factors in Table 2 were performed. After adjustment there was a significant independent increased risk of not being breast-fed at three months in relationship to maternal age (<25 years OR 4.2, 95% CL 2.0-8.5), maternal smoking (OR 4.0, 95% CL 2.1-7.4), neonatal feeding (supplements OR 3.9,95% CL 2.1-7.2) and initial weight loss ≥10% OR 2.8, 95% CL 1.4-5.8).

 

Other Findings

The relative risk of not being breast fed at three months in relation to feeding in the maternity unit and initial weight loss(reference group=those exclusively breast fed with an initial weight loss of less than 10% of birth weight, OR 1.0) was analyzed.

There was an interaction between initial weight loss and feeding pattern at the maternity unit and the risk of not being breast-fed at three months. The risk of not being breast-fed was almost seven times higher for infants who were supplemented and had an initial weight loss >=10% of birth weight, compared with unsupplemented infants with less weight loss.

• Giving supplementary donor’s milk or formula during early neonatal period increased the risk of a short duration of breastfeeding even after adjustment for initial weight loss and other factors examined
• Supplementary feeding in the maternity unit was associated with an increased risk of early cessation of breastfeeding. The authors recommended that supplementary feeding be given with greater restrictively to healthy newborn children.

 

Limitations

The question explored by the researchers was whether the association shown between supplementation in the maternity unit and a shorter duration of breast feeding was casual or spurious. A number of confounding factors were considered in the analysis, characterizing the mother and child.

• Maternal education which has been shown to be associated with breastfeeding was not included in the analysis; smoking, a factor strongly associated with education was examined
• The association between supplementation and non-breastfeeding was not weakened but became stronger when the different potential confounders were accounted for. This may have indicated a casual relationship; however, the mother’s confidence and willingness to continue breastfeeding was not assessed.
• The supplements were planned to be prescribed by pediatricians but some children received supplementation for no declared reason
• The routines for supplementation were rather liberal and were later revised in accordance with international recommendations.

 

Analytical longitudinal surveys refer to what epidemiologists term prospective or cohort studies. A Cohort Study is a study in which patients who presently have a certain condition and/or receive a particular treatment are followed over time and compared with another group who are not affected by the condition under investigation. Studies of this kind provide a better opportunity than one-time cross-sectional studies to examine whether certain behaviors do in fact lead to (or cause) the disease.            

The limitations and critique of the study, as stated by the authors appear to be very appropriate.