Full Milk Feeds From Day 1 in Late Preterm Infants: Large RCT Shows No Shorter Hospital Stay and No Signal for Increased NEC or Hypoglycaemia

Highlight

Full milk feeds from the first day of life in clinically stable infants born at 30+0 to 32+6 weeks did not reduce length of hospital stay compared with gradual feeding supplemented by IV fluids or parenteral nutrition. There was no increase in necrotising enterocolitis or severe hypoglycaemia; serious adverse events were uncommon and unrelated to allocation.

Background

Late preterm infants (commonly defined as 34+0 to 36+6 weeks) and those born slightly earlier in the moderate preterm window (30+0 to 32+6 weeks) make up a substantial proportion of preterm births. Feeding management in preterm infants balances two priorities: establishing enteral nutrition and avoiding complications such as necrotising enterocolitis (NEC) and hypoglycaemia. The conventional approach—gradual advancement of enteral feeds supported by intravenous fluids or parenteral nutrition—aims to protect the immature gut but prolongs exposure to IV lines and hospital-based care.

Proponents of early, full enteral feeding argue it could shorten hospital stay, reduce the duration of parenteral nutrition and associated line-related infections, and promote breastfeeding establishment. Critics remain concerned about the potential to precipitate NEC or hypoglycaemia in vulnerable infants. High-quality randomized evidence is needed to inform practice for the specific gestational-age window of 30+0 to 32+6 weeks.

Study design

Trial overview

FEED1 was an open-label, parallel-group, multicentre, randomized, superiority trial conducted in 46 neonatal units across the UK. The trial enrolled mothers of infants born between 30+0 and 32+6 weeks’ gestation. Recruitment occurred within 3 hours of birth; infants had to be clinically stable. Key exclusions included congenital anomalies that contraindicated enteral feeds and small-for-gestational-age infants with reversed end-diastolic flow on antenatal Doppler.

Interventions and comparators

Participants were randomized via a web-based minimisation algorithm (with a random element) to one of two arms:

– Full milk feeds from day 1: 60–80 mL/kg per day of milk (enteral-only fluids).
– Gradual milk feeding: a maximum of 30 mL/kg per day on day 1, with intravenous fluids or parenteral nutrition to meet fluid and nutritional needs.

This pragmatic design left certain clinical details (type of milk, feeding method) to local policies and clinicians, reflecting real-world practice. The study was open-label for clinicians and families; investigators and statisticians were masked until database lock.

Primary and safety outcomes

The primary outcome was length of hospital stay. Prespecified safety outcomes included NEC and events of hypoglycaemia. Analysis used the intention-to-treat principle. The trial was prospectively registered (ISRCTN89654042) and funded by the UK NIHR; longer-term follow-up to 24 months was planned.

Key findings

FEED1 enrolled 1761 mothers with 2088 infants (1047 allocated to full milk feeds; 1041 to gradual feeding). Mean gestational age was 31.7 weeks in both groups; mean birthweights were similar (approximately 1.62 kg).

Primary outcome

– Length of hospital stay: mean 32.4 days (SD 13.3) in the full milk group vs 32.1 days (SD 13.5) in the gradual feeding group.
– Adjusted difference between means: -0.02 days (95% CI -1.07 to 1.03); p = 0.97.

Interpretation: There was no evidence that initiating full enteral feeds from day 1 reduced duration of inpatient care in this gestational-age group.

Safety and secondary outcomes

– Survival to discharge: 99.6% in both groups (1030/1034 vs 1027/1031).
– Necrotising enterocolitis: 4/1030 (0.4%) in the full milk group vs 6/1027 (0.6%) in the gradual feeding group (absolute difference -0.19 per 100 infants; 95% CI -0.80 to 0.41).
– Hypoglycaemia (blood glucose <2.2 mmol/L): mean number of tests showing this value 0.6 (SD 1.0) vs 0.5 (SD 0.7), essentially similar between groups.
– Serious adverse events: uncommon and balanced (8 [0.8%] vs 10 [1.0%]); reported events were judged unrelated to trial allocation.

Notable features

– Missing primary outcome data were balanced (18 infants in each arm).
– The overall incidence of NEC was low across both arms.

Clinical and statistical significance

The trial was large and well-conducted for the chosen primary outcome. The near-zero adjusted difference and narrow confidence interval indicate a high degree of precision around the null effect for length of stay in this population. Importantly, the absence of an observable excess risk of NEC or hypoglycaemia with immediate full enteral feeding is clinically reassuring, although the absolute rates of NEC were low, limiting power for rare events.

Expert commentary and interpretation

FEED1 provides robust, pragmatic randomized evidence specific to infants born between 30+0 and 32+6 weeks. The trial’s strengths include its large sample size, multicentre recruitment, early randomization (within 3 hours of birth), intention-to-treat analysis, and masked data handling.

Limitations to consider:

– Open-label design: Clinicians and parents were unblinded, which could theoretically influence management decisions after randomization (for example, timing of discharge). However, investigators masked to allocation until analysis reduces analytic bias.
– Restricted population: Results apply to clinically stable infants in the 30+0 to 32+6-week range. Findings should not be extrapolated to more preterm infants (<30 weeks), very small‑for‑gestational‑age infants, or those who are clinically unstable at birth.
– NEC is rare in this cohort: Low overall NEC rates mean the trial may lack precision to detect small (but potentially important) differences in NEC incidence.
– Primary outcome (length of stay) is multifactorial: Many non-nutritional factors influence discharge timing (respiratory stability, thermoregulation, social factors), which could dilute a potential feeding effect.

Mechanistic plausibility and clinical context

Early full enteral feeding removes intravenous access earlier and may theoretically reduce catheter-related infections and the metabolic instability associated with parenteral nutrition. It may also support breastfeeding by providing more opportunities for enteral feeding and earlier milk expression by mothers. Conversely, concerns that a larger early enteral volume might overwhelm immature gut perfusion and bacterial colonisation have historically driven conservative approaches. FEED1 suggests that in this gestational band, early full enteral feeds do not increase the recognized clinical harms.

Clinical implications and practice recommendations

For clinicians caring for infants born between 30+0 and 32+6 weeks who are clinically stable at birth, FEED1 supports the safety of initiating full enteral feeds from day 1 with respect to NEC and hypoglycaemia. However, full enteral feeds did not shorten hospital stay in this trial, which tempers expectations about discharge gains from this strategy alone.

Suggested pragmatic approach:

– Consider offering full enteral feeding from day 1 for clinically stable infants in this gestational group after shared discussion with parents and neonatal team, especially where local practice already supports early feeding.
– Maintain vigilant monitoring for feeding intolerance and hypoglycaemia, as per unit protocols.
– Individualize decisions for infants with antenatal or perinatal risk factors (e.g., SGA with abnormal Dopplers, persistent cardiorespiratory instability), for whom the trial excluded or did not assess these subgroups.
– Continue efforts to support maternal expression and lactation, as earlier enteral feeding may facilitate breastfeeding establishment.

Resource and policy implications

Wider adoption of immediate full enteral feeding might reduce use of intravenous fluids and parenteral nutrition resources and associated line care burdens. However, because length of stay was not reduced in FEED1, policymakers should be cautious in assuming hospital bed-day savings from changing feeding protocols alone.

Research gaps and future directions

– Lower gestational ages: High-quality trials are needed in infants <30 weeks, who have higher NEC risk and different metabolic needs.
– Longer-term outcomes: The ongoing 24-month follow-up from FEED1 will be important to assess growth and neurodevelopmental outcomes related to early feeding strategies.
– Subgroups: Better characterization of outcomes in SGA infants, infants receiving donor human milk versus formula, and those with intrauterine growth restriction is required.
– Mechanistic studies: Research into gut microbiome changes, intestinal perfusion, and biomarkers of feeding tolerance could clarify biological effects of early full feeds.

Conclusion

In clinically stable infants born at 30+0 to 32+6 weeks’ gestation, initiating full enteral milk feeds from day 1 did not reduce length of hospital stay compared with a gradual feeding approach supplemented by intravenous fluids or parenteral nutrition. Importantly, early full feeds were not associated with an increased risk of necrotising enterocolitis or clinically important hypoglycaemia in this population. These findings support the safety of immediate full enteral feeding for selected late to moderate preterm infants, while emphasizing individualized clinical judgment and the need for further data in more preterm and high-risk groups.

Funding and trial registration

FEED1 was funded by the UK National Institute for Health and Care Research (NIHR). The trial was prospectively registered: ISRCTN89654042.

References

Ojha S, Mitchell EJ, Johnson MJ, Gale C, McGuire W, Oddie S, Hall SS, Meakin G, Anderson J, Partlet C, Su Y, Johnson S, Walker KF, Ogollah R, Mistry H, Naghdi S, Montgomery A, Dorling J; FEED1 collaborative. Full exclusively enteral fluids from day 1 versus gradual feeding in preterm infants (FEED1): a open-label, parallel-group, multicentre, randomised, superiority trial. Lancet Child Adolesc Health. 2025 Dec;9(12):827-836. doi: 10.1016/S2352-4642(25)00271-8. Epub 2025 Oct 17. PMID: 41115446.

(Additional guideline and systematic review summaries are available on request.)