Our study's objectives were to ascertain when TF was started and its determinants in preterm newborns hospitalized to Yabello General Hospital and Bule Hora University Teaching Hospital's NICU between January 1, 2022, and September 30, 2022. As a result, of the 411 neonates observed on the first day of follow-up, 21.93% began TF, and 28 were censored. However, only 24 were left at the end of the five-day follow-up, with 14 of them having started TF. The overall incidence density in our sample was 43.6 per 100 neonate days. This means, out of every 100 days that newborn babies were observed in our study, TF was started in an average of 43.6 of those days. The median (IQR) time to TF onset of was 2 (1–4) days. Moreover, he study revealed a significant increase in preterm neonatal mortality in Ethiopia, with 113.04 deaths per 1000 neonate-days, a significant increase from previous research. GA (> 34 weeks), SVD, the absence of sepsis, and the absence of hypothermia increase the probability of early initiation of TF by 64%, 76%, and 51%, respectively. Conversely, it was discovered that delayed TF start was caused by BW (< 1500 kg), lack of kangaroo mother care, and being admitted to a level-2 hospital.
In our study, we found that only 21.93% of preterm neonates initiated feeding within the first 24 hours of birth, indicating that many did not start within this optimal timeframe. The study's findings are lower than previous studies in Portugal 44% (40), Tuscany Italy (74.1%) (41), Iran (36%) (42), Nigeria and Kenya 48% (28), New Zealand (60%) (23), Addis Ababa (29.9%) (32) and Amhara, Northwest Ethiopia (24.76%) (19). The discrepancy may be due to differences in service quality, and availability of specialized facilities, milk banks, study period length, sample size, and design between the studies. Moreover, nearly half, (46.74%) of preterm neonates initiated TF feeding within 48 hours of birth. This is comparable with a study conducted in Addis Ababa, Ethiopia (48.2%) (32). However, it is lower than the findings from Nigerian Special Care Baby Unit (66.7%) (43), Iran (63.2%) (42), Uganda rural hospitals (80%) (30) and New Zealand (80%) (23). This highlights a significant delay in initiating TF among preterm neonates and creates concerns about high neonatal mortality as evidenced by our study where, over 50% of the neonate died within the first 48 hours of life. This could be because of their initial treacherous condition. The result of this study regarding incidence density of TF is lower than the result of a follow up study conducted in northern Ethiopia (19), where the incidence density of starting TF was 48 per 100 neonate-days. Moreover, the median (IQR) time of starting TF found in our study 2 (IQR: 1–4) days is slightly higher than the result of a follow up study conducted in China (44) and a retrospective study in New Zealand (23) which had a similar median time of 1 day. This discrepancy might be due to difference in service delivery quality, hospital level, having well-equipped institution, sample size and study settings. However, consistent with a longitudinal study in Ethiopia in which median time to start TF was less than 2 days (19) and a multicenter observational follow up study in Nigeria and Kenya (35), which might be due to similarity in study design and characteristics of study subjects. This implies that the presence of a significant delay to initiate TF for studied preterm neonates.
Preterm neonatal mortality (NMR) was found to be surprisingly common in our study. We found an astounding 113.04 deaths per 1000 neonate-days, a significant increase above earlier research done in various parts of Ethiopia. From 27 per 1000 neonate-days in southern Ethiopia (45), 29.438 per 1000 neonate-days in Addis Ababa, Central Ethiopia (46) and to 75.63 per 1000 neonate-days in northwestern Ethiopia (47), these previous investigations revealed substantially lower fatality rates. The NMR of 45.15 per 1000 neonate-days was reported in a study conducted in Hawassa City, Ethiopia (25), which is geographically similar to our study area. This is noteworthy. Even the neighboring African country, Burkina Faso had a much lower rate, 1.93 deaths per 1000 person-days (48). The study's unexpected high NMR result led to a potential explanation of a low proportion of neonates initiating TF early in their lives. Previous research suggests that preterm neonates who don't receive TF early are more susceptible to death (46). Studies have shown that early TF initiation offers several benefits – it reduces the length of hospital stays, lowers infection rates, and promotes gut development (22). Conversely, a delay in starting TF might hinder the maturation of the gut, leading to a weakened immune system (49, 50). This, in turn, could significantly increase the risk of complications like Necrotizing Enterocolitis (NEC) and various infections, ultimately contributing to higher neonatal mortality (51). Therefore, we strongly suggest that, in particular, for susceptible groups such as preterm newborns, targeted interventions encouraging early TF initiation could drastically lower NMR rates and increase survival rates.
In this study, the chance of starting TF on time was 55% reduced among a very low birth weight (< 1500gm) preterm neonates when compared to those with a birth weight of > 1500gm. This finding is in line with the results of retrospective longitudinal a study in China (44), a retrospective study in New Zealand (23) and a multicenter prospective study conducted in Ethiopia (19). This might be due to perceived under-development of organs and low readiness to enteral feeding of preterm neonates with lower birth weight. However in contrast to this, guidelines on feeding of very low birth weight infants recommend early initiation of TF within 1 day after birth, also for this group of preterm neonates (9). This result highlights the importance of individualized care plans based on birth weight in order to insure early TF for VLBW preterm neonates.
Among preterm neonates with a gestational age of less than 34 weeks, there was a statistically significant 39% reduction in the likelihood of initiating TF in comparison to neonates with a gestational age of 34 weeks or more. This result is consistent with findings of observational prospective follow up study in Italy (41), retrospective follow up study in China (44), a cross sectional study carried out in two of African countries (28), a multicenter prospective study conducted in Ethiopia (18), prospective follow up study conducted in northern Ethiopia (19) and cross sectional study in Spain (15), which revealed delayed commencement of TF among newborns with lower gestational age. This might be due to lack of intestinal maturation, anxiety of feeding resistance, and perceived danger of NEC, preterm neonates born before 34 weeks of pregnancy may have gastrointestinal issues. This implies that early identification and support of lower gestational age preterm neonates may improve the time to initiate TF among this group of preterm neonates.
The study found that preterm neonates delivered through spontaneous vaginal delivery had a 64% higher likelihood of starting TF compared to those delivered through CS, possibly due to delayed breast milk provision. The study's findings align with a multicenter prospective follow-up conducted in northern Ethiopia (19), prospective follow-up conducted in Addis Ababa, Ethiopia (32), and Italy(27). This might be because of the reason that newborn's post-delivery needs may require additional monitoring and assistance, while the mother's recovery from the surgery may take longer. The result implies that preterm neonates delivered with CS are at risk of experiencing delayed initiation of TF and related complications. Thus, the healthcare providers must closely monitor both mothers and babies post-delivery and initiate enteral feeding as soon as the baby is stable and ready. However, this finding is in contrast with a result of observational retrospective follow up study conducted in China which revealed lack of significant association of delivery mode and time to initiate enteral feeding (44). This discrepancy might be due to difference in characteristics of studied subjects, sample size and study setup.
WHO recommended KMC for preterm or low-birth-weight infants, starting in healthcare facilities or at home, and lasting 8–24 hours daily (52). In line with this evidence, in our study, absence of KMC was associated with a 41% reduced likelihood of earlier TF initiation as compared to those with the care. Similar to this finding, a retrospective follow up study in Turkiye has identified KMC as a main factor to improve enteral feeding skills among preterm neonates (53). Moreover, a prospective cohort study done in a teaching hospital in India identified that early KMC was safe and associated with reduced time to full feeds (TFF) in preterm neonates (54). A meta-analysis of randomized controlled trial studies reported that KMC encourages early breastfeeding initiation among preterm and low birth weight infants (55). This implies that KMC is a gentle, effective method for preterm infants, promoting their health and well-being by allowing early discharge and avoiding agitation in busy wards (56). On the other hand, preterm neonates without hypothermia had a 51% higher hazard of starting TF when compared to those with hypothermia. This is line with a study conducted in Kuala Lumpur Maternity Hospital (57). This might be due to the fact that most preterm neonates with hypothermia often receive radiant warmer treatment and stay separated from mothers, making early enteral feeding unsuitable. This implies the significance of effective temperature regulation and thermal management strategies in the care of preterm neonates to facilitate timely initiation of TF.
In this study, the hazard of starting TF was 76% increased among premature neonates without sepsis as compared to those with sepsis. This finding is supported with a result of a retrospective study conducted in Kaplan medical center, Israel (58), Addis Ababa, Ethiopia (32), and Maharashtra, India (59). This similarity might be due to the fact that sepsis can cause decreased gut motility, increased risk of NEC (60), and damage the gut lining (61), making it difficult for neonates to tolerate enteral feeding. Thus, the healthcare providers may delay TF initiation to control infection, reduce NEC risk, and minimize gut permeability. However in contrast to this, guideline on TF does not consider sepsis as contraindication of TF among premature infants and recommends enteral feeding to be initiated early for this group of preterm infants (9). The result implies early prevention and treatment of sepsis among preterm neonates might support timely initiation of TF.
Our study revealed inconsistency in how TF is initiated for preterm neonates across different healthcare levels. We observed a surprising trend – neonates admitted to the secondary-level hospital (YGH) were 22% less likely to receive TF compared to those in the tertiary hospital (BHTRH). This finding is particularly concerning, as timely initiation of TF is crucial for the survival and development of preterm infants. Similar trends were recognized in prior research conducted in Addis Ababa, Ethiopia (32), highlighting a potential issue within the Ethiopian healthcare system. Additionally, studies across geographically distinct locations like Nigeria and Kenya have reported similar disparities (28, 35). This observed similarities suggests a concerning possibility that preterm neonates in less-resourced facilities may face systemic disadvantages when it comes to accessing essential and life-saving interventions like TF. Moreover, the secondary hospitals might have lack of resources, trained personnel, and logistical challenges. Therefore, our study highlights the need for a comprehensive investigation into the disparity in access to TF across healthcare levels, including capacity building, resource allocation, and logistical processes.
To the best of our knowledge our study is among few in Ethiopia conducted to determine the time to initiate trophic feeding, involving advanced statistical analysis. However, due to retrospective nature, incomplete data was excluded, potentially introducing selection bias. The data was collected from secondary sources, potentially missing important predictors like nurse-patient ratio and breast milk availability.