Table 3.
Human milk (HM) and developmental outcome in very low birth weight (VLBW) infants: perinatal, postnatal, social and environmental data*
| Reference | Perinatal/postnatal factors | Social and environmental factors |
|---|---|---|
| Pinelli et al. [29] | Type of delivery | Maternal and paternal age, education and occupation, 1 or 2 parent home, social classes I-V (Hollingshead index) |
| Furman et al. [30] | Delivery at perinatal center, antenatal steroid and cesarean section. Apnea, sepsis, jaundice, necrotizing enterocolitis, chronic lung disease, cranial ultrasound abnormalities. | Maternal education and ethnicity, and marital status |
| Tanaka et al. [31] | Chronic lung disease, cranial ultrasound, necrotizing entercolitis. Intrauterine growth retardation | Maternal age and education |
| Horwood et al. [32] | Sex, multiple births, birth weight, gestational age, intrauterine growth retardation, 5 min Apgar score | Maternal age, education and smoking, 1 or 2 parents, family income, child ethnicity, birth order |
| Smith et al. [33] | Length of hospital stay | Maternal age, verbal ability, education, cigarette smoking and marital status, Home observation for measurement of the environment inventory – short version, annual household income, gender, parity |
| Johnson et al. [34] | Birth weight, gestation, antenatal steroid, premature rupture of membranes, vaginal breech delivery, chorioamnionitis, admission temperature <35°C, CRIB score, abnormal last cranial ultrasound, necrotizing entercolitis, postnatal steroid, duration of NICU admission. Neurodevelopmental assessment results at 30 m and 6 y | Socioeconomic (UK National Statistics Socio-Economic classification), maternal age, race and highest education. |
| Vohr et al. [35] | Gestation, gender, sepsis, intraventricular hemorrhage grade 3 to 4, periventricular leukomalacia, oxygen need at 36 weeks, necrotizing enterocolitis, and weight <10th percentile at 18 months. | Maternal age and education, marital and health insurance status, race, and income. |
*The variables entered into the final model to determine the independent effect of HM feeding were varied and not always fully described. Some investigators [30] used composite scores to minimize the number of variables entered into data analysis and no specific modeling was performed by other investigators [31].