794 - Time to First Human Milk Feed Among Premature Neonates and Relationship with Clinical Outcomes in a Tertiary Referral Hospital in Addis Ababa, Ethiopia

Publication Number: 794.6630

Gregory C.. Valentine, University of Washington School of Medicine, Tacoma, WA, United States; Krystle M.. Perez, University of Washington, Seattle, WA, United States; Olivia C. Brandon, University of Washington: Magnuson Health Sciences Center – RR 544A, Seattle, WA, United States; Gal Barbut, Golisano Children's Hospital at The University of Rochester Medical Center, Rochester, NY, United States; Sharla Rent, Duke University School of Medicine, Durham, NC, United States; Merhawit Abadi, Saint Paul's Hospital Millennium Medical College, ADDIS ABABA, Adis Abeba, Ethiopia; Redeat Workneh, St Paul's Hospital Millennium Medical College, Addis Ababa, Ethiopia, Adis Abeba, Ethiopia; Mahlet Abayneh, St Paul's Hospital Millennium Medical College, Addis ABABA, Adis Abeba, Ethiopia

Background: The World Health Organization recommends feeding neonates within one hour after birth. In resource-limited settings, it is uncertain how the prioritization of early feeding may impact other clinical care and support for vulnerable small and/or sick newborns.

Objective: Among premature neonates admitted to the St. Paul’s Hospital Millennium Medical College (SPHMMC) tertiary referral hospital neonatal intensive care unit (NICU) in Ethiopia, we sought to evaluate the relationship between time of first enteral feed with all-cause mortality, length of hospital stay (LOS), and a composite outcome of culture-positive sepsis, suspected necrotizing enterocolitis (NEC), and mortality.

Design/Methods: We performed a retrospective, observational, cross-sectional study including premature neonates ( < 37 weeks) admitted to the SPHMMC NICU from Jan-May 2024. Neonates were excluded if they had: (a) known congenital anomalies, (b) a length of stay of < 2 days or (c) had died within 2 days after birth. For length of stay analyses we assessed: (a) only survivors, and (b) imputing 60 days as the length of stay to include those that died. Time to first feed was categorized into < 1 hour, 1- < 12 hours, 12- < 24 hours or ≥24 hours from birth. Categorical and continuous outcomes were assessed using logistic and linear regression models, respectively.

Results: 183 premature neonates were included in this analysis. Table 1 reports maternal and neonatal characteristics of feeding groups. When compared to feeding < 1 hour after birth, neonates fed between 1- < 24 hours (combined groups due to 0% mortality in 1- < 12 hour group) had significantly reduced odds of mortality in unadjusted analyses (4.6% vs 21.4%, Odds Ratio [OR] 0.18, 95% Confidence Interval [CI] 0.03, 0.99) but not within the adjusted model (adjusted OR [aOR] 0.09, 95% CI 0.01, 1.20; Figure 1). Neonates fed between 1- < 12 hour had significantly fewer hospital days than those fed < 1 hour after birth (aOR -11.40 days, 95% CI -20.70, -2.08; Table 2). Compared to feeding < 1 hour after birth, the lowest odds of the composite outcome occurred in the 1- < 12 hour feeding group (5.5% vs 35.7%, aOR 0.10, 95% CI 0.01, 1.14, p=0.06)

Conclusion(s): Initiation of enteral feeds within 1- < 12 hours may be associated with improved clinical outcomes as compared to initiation of feeds within 1 hour after birth. Educational, advocacy, and implementation efforts are needed to highlight the importance of timely initiation of enteral feeds while also promoting and ensuring clinical support to these small and/or sick newborns.

Table 1: Time to Receive First Human Milk Feed and Maternal and Neonatal Characteristics Between Groups
NEC: Necrotizing enterocolitis
All data presented as n/N (%), median ± IQR. Chi-Square test or ANOVA for categorical and continuous variables, respectively. Only infants surviving at least 2 days included. Significant p-values bolded. Outcomes analyses are unadjusted. See additional tables and figures for adjusted analyses using linear and logistic regression for continuous and categorical outcomes, respectively.
& Due to 0 cases of mortality within the 1- <12 hour group, unable to compute p-value. See Figure 3 for separate analyses on mortality. P-values
# In the Length of Stay analyses, there were 3/14, 3/36, 6/29, 13/104 with missing dates in each of the four groups ( <1 hour, 1- <12 hours, 12- <24 hours, ≥24 hours), respectively, that were not included in these analyses and including cases of mortality or missing dates of birth or discharge.
* Imputed 60 days for any case of mortality (1 more than the longest length of stay of a surviving baby), there were 1/14, 1/36, 3/29, 0/104 with missing dates in each of the four groups ( <1 hour, 1- <12 hours, 12- <24 hours, ≥24 hours), respectively, that were not included in these analyses and including missing dates of birth or discharge.


Table 2: Unadjusted and Adjusted Logistic Regression Models Evaluating Time to First Feed and Clinical Outcomes
#Covariates included within the adjusted generalized linear models included significant variables in Table 1 with p values of 0.10 or less: gestational age, birthweight, antenatal corticosteroids, inborn status, maternal age, and pre-eclampsia
&Composite outcome included suspected necrotizing enterocolitis, culture-positive sepsis, and all-cause mortality. Of note, due to no cases of suspected NEC or culture-positive sepsis occurring within the <1 hour and the 1- <12 hour feeding groups, we only evaluated these outcomes as a composite with mortality.


Figure 1: Evaluation of Time to First Feed and All-Cause Mortality.
(A) Represents the unadjusted and adjusted analyses of all-cause mortality and time to first feed category comparing the following 3 categories: <1 hour, 1- <24 hours, and ≥24 hours after birth. Covariates included within the adjusted model include: gestational age, birthweight, antenatal corticosteroids, inborn status, maternal age, and maternal pre-eclampsia. (B) Represents the unadjusted mortality proportion for 12-hour categories. Of note, the 1- <12 hour category did not have any events of mortality that occurred, which is why we needed to subsequently combine the 1- <12 hour and 12- <24 hour categories into a 1- <24 hour category as seen in panel C.