Menu
Menu
Menu
Menu
Household food insecurity has been hypothesized to negatively impact breastfeeding practices and breast milk intake, but this relationship has not been rigorously assessed. To generate an evidence base for breastfeeding recommendations among food-insecure mothers in settings where HIV is highly prevalent, we explored infant feeding practices among 119 mother–infant dyads in western Kenya at 6 and 24 weeks postpartum. We used the deuterium oxide dose-to-the-mother technique to determine if breastfeeding was exclusive in the prior 2 weeks, and to quantify breast milk intake. Sociodemographic data were collected at baseline and household food insecurity was measured at each time point using the Household Food Insecurity Access Scale. Average breast milk intake significantly increased from 721.3 g/day at 6 weeks postpartum to 961.1 g/day at 24 weeks postpartum. Household food insecurity at 6 or 24 weeks postpartum was not associated with maternal recall of exclusive breastfeeding (EBF) in the prior 24 hr or deuterium oxide-measured EBF in the prior 2 weeks at a significance level of 0.2 in bivariate models. In a fixed-effects model of quantity of breast milk intake across time, deuterium oxide-measured EBF in the prior 2 weeks was associated with greater breast milk intake (126.1 ± 40.5 g/day) and every one-point increase in food insecurity score was associated with a 5.6 (±2.2)-g/day decrease in breast milk intake. Given the nutritional and physical health risks of suboptimal feeding, public health practitioners should screen for and integrate programs that reduce food insecurity in order to increase breast milk intake.
Data were drawn from a prospective cohort study that was designed to compare differences in breast milk intake between HIV‐uninfected infants whose mothers were either HIV‐infected or ‐uninfected at enrolment (PACTR201807163544658). Recruitment criteria and the study design have been described elsewhere (Oiye, Mwanda, Mugambi, Filteau, & Owino, 2017). Briefly, mothers (n = 143) attending the Maternal and Child Health Clinic of Siaya County Referral Hospital in western Kenya for infant vaccinations at 6‐weeks postpartum were recruited into the study between February and September 2014. Women were eligible for inclusion if they planned to live in Siaya District for the following 10 months and their infants were both 6 weeks of age (±8 days) and HIV‐uninfected at enrolment. Women were excluded if their infants had low birthweight (<2,500 g), were born preterm, and/or were unable to breastfeed. Women were systematically sampled to detect differences in breast milk intake by maternal HIV status at a power of 0.8 (i.e., women living with HIV were oversampled), allowing for an expected 33% loss to follow‐up. This was an appropriate setting to examine the impacts of food insecurity and HIV on breast milk intake because 34.0% of households in the region were food insecure (Siaya County Integrated Development Plan 2013–2017, 2014). Additionally, at the time of study enrolment, the prevalence of HIV among women of reproductive age in Siaya County was nearly four times the national average (Kenya Ministry of Health, 2014). Survey data were collected at 6 and 24 weeks postpartum by local clinic‐based study nurses using a structured paper questionnaire. Sociodemographic data (e.g., age, highest level of education), delivery information, and infant feeding practices were based on maternal recall. Following WHO guidelines, EBF was defined as an infant solely receiving breast milk, except for medically prescribed oral rehydration salts, drops, and/or syrups (vitamins, minerals, medicines) (World Health Organization, 2008). Date of birth and infant birthweight were obtained from both mother and child clinic cards when available, otherwise they were obtained by maternal recall. Anthropometric measurements were taken following standard protocols, as described elsewhere (Oiye et al., 2017). Household food insecurity was measured using the Household Food Insecurity Access Scale (range: 0–27; Coates, Swindale, & Bilinsky, 2007), which prompts about experiences of food insecurity in the past 4 weeks. A subset of these items was used to measure household hunger (range: 0–6; Ballard, Coates, Swindale, & Deitchler, 2011). Households were then classified as having low (0–1), moderate (2–3), or high (4–6) hunger. Maternal HIV status was assessed at 6 and 24 weeks postpartum using colloidal gold antibody tests (KHB Shanghai Kehua Bioengineering Co. Ltd). At 6 weeks of age, each infant's serostatus was assessed with HIV‐1 DNA polymerase chain reaction (PCR) using a T100 Thermal Cycler (Bio‐Rad Laboratories Inc, UK). Infant HIV status was not assessed at 24 weeks because Kenyan government guidelines require HIV‐exposed infants to be tested at 9 months of age (Kenya Ministry of Health, 2012). On the basis of antibody and PCR‐confirmatory tests at 9 months, two infants seroconverted and were therefore excluded from analysis. Breast milk intake was measured at 6 and 24 weeks postpartum using the DTM technique, as described by the International Atomic Energy Agency (IAEA) (International Atomic Energy Agency, 2010). This method for measuring breast milk intake has been shown to be effective in similar East African settings (Ettyang, van Marken Lichtenbelt, Esamai, Saris, & Westerterp, 2005; Getahun et al., 2017). After being weighed, baseline (pre‐dose) saliva samples were collected from the mother and the infant (day 0) using a syringe and sterile cotton wool. Mothers then consumed a 30‐g oral dose of deuterium oxide (D2O) through a straw and were instructed to feed their infant as usual. At least 2 ml of saliva samples were subsequently collected from both the mother and infant in 10‐ml polypropylene sterile tubes on days 1, 2, 3, 4, 13, and 14. Post‐dose saliva was collected at the same time of day as at baseline collection. All samples were separately secured in ziplocked polythene bags and immediately frozen in −20°C freezers. Samples were then transported in iceboxes to the Kenya Medical Research Institute (KEMRI) Nutrition Laboratory in Nairobi for analysis. Deuterium enrichment in mother and infant saliva over a 14‐day period was measured using a Fourier Transform Infrared Spectrophotometer (FTIR 8400 Series; Shimadzu Corporation, Kyoto, Japan). Deuterium enrichment was in turn used to calculate maternal total body water, from which breast milk output was derived using computer modelling (International Atomic Energy Agency, 2010). Infant intake of water from sources other than breast milk was calculated using a spreadsheet developed by the IAEA (International Atomic Energy Agency, 2010). Infants were considered EBF using the DTM technique if they consumed <25 g of water from sources other than human milk per day, as recommended by the IAEA (International Atomic Energy Agency, 2010). No mothers reported giving oral rehydration salts to their infants at either time point, such that this was an appropriate cut‐off. Survey and biomarker data were entered into Epi Info Version 6 and cleaned using both Microsoft Excel and Stata 14.0 (StataCorp, College Station, TX, USA). Descriptive statistics (chi‐square, t tests) were performed using Stata 14.0 with a significance level of .05. Significant covariates (p < .2) of EBF or human milk intake in bivariate analyses were included in multivariable linear regressions for each time point; variables were eliminated using a backwards stepwise approach (p < .1). Model specification was evaluated using the link test. To examine the impact of food insecurity on breast milk intake over time, a fixed effects model was built using a similar stepwise technique. Given our restricted sample size, a bootstrap draw of 5,000 was used to check the asymptotic normality assumption, make accurate inferences about the broader population, and develop robust confidence intervals (Carpenter & Bithell, 2000). All models of breast milk intake were adjusted for deuterium‐oxide measured EBF. The Kenyatta National Hospital/University of Nairobi Ethics and Research Committee (KNH/UON/ERC) approved this study. The approval reference number is KNH‐ERC/A/90 (P517/09/2012). All participants provided written informed consent.
N/A
