Short versus long feeding interval for bolus feedings in very preterm infants

Review question

Are short feeding intervals (for example, two hours or shorter) better tolerated than long feeding intervals (three hours or longer) for regular milk feeds in very preterm infants?


Feeding very preterm infants, less than 32 weeks' gestation at birth, is a major challenge. They have an immature gut which may lead to problems, varying from mild (feeding intolerance) to moderate (regurgitation of milk from the stomach) to severe (such as necrotising enterocolitis; NEC). NEC is an infectious complication leading to irreparable loss of parts of the bowel. The feeding interval, that is, the time interval between each feed, might matter but determining the appropriate feeding interval is a major challenge. Both short intervals, typically less than three hours, and longer intervals of three or more hours have their own risks. When the interval is short, a smaller volume of milk can be given more frequently. The infant might tolerate smaller volumes better, but their gut might not have sufficient time to rest between each feeding.

Study characteristics

We searched medical databases up 25 June 2020. We found four studies (involving 417 infants), all conducted in middle-income countries. All four studies compared two-hourly with three-hourly feeding interval. All studies involved very low birth weight infants with a gestational age (time since the beginning of the woman's last menstrual period) range from 29 weeks to 35 weeks.

Key results

When we combined the results of the studies, they suggested little or no difference between the two- and three-hourly feeding intervals. We are uncertain whether there is any difference between two- and three-hourly feeding in terms of days to achieve full enteral (tube) feeding because the results were imprecise and there were biases in the studies. Days taken to regain the birth weight, after the usual initial drop in weight during the first days after birth, may be slightly longer in infants receiving two-hourly feeds, but we are uncertain of the importance of this. We did not have enough data to determine whether there was any difference in any of the adverse outcomes that the studies reported. However, because only a small number of infants experienced NEC, we are uncertain whether there is any difference between the groups. There was no information on the effect of feeding interval on death, infant growth during hospital stay and neurodevelopmental outcome (brain development in childhood), and there was no information about other feeding intervals such as one-hourly or four-hourly feeds. There is one study awaiting classification as we need more information.

Quality of evidence

The quality of the evidence in this review was low. Therefore, there may be no clinically important differences between two- and three-hourly feeding intervals.

Authors' conclusions: 

The low-certainty evidence we found in this review suggests that there may be no clinically important differences between two- and three-hourly feeding intervals. There is insufficient information about potential feeding complications and in particular NEC. No studies have looked at the effect of other feeding intervals and there is no long-term data on neurodevelopment or growth.

Read the full abstract...

There is presently no certainty about the ideal feeding intervals for preterm infants. Shorter feeding intervals of, for example, two hours, have the theoretical advantage of allowing smaller volumes of milk. This may have the potential to reduce the incidence and severity of gastro-oesophageal reflux. Longer feeding intervals have the theoretical advantage of allowing more gastric emptying between two feeds. This potentially provides periods of rest (and thus less hyperaemia) for an immature digestive tract.


To determine the safety of shorter feeding intervals (two hours or shorter) versus longer feeding intervals (three hours or more) and to compare the effects in terms of days taken to regain birth weight and to achieve full feeding.

Search strategy: 

We used the standard search strategy of Cochrane Neonatal to run comprehensive searches in CENTRAL (2020, Issue 6) and Ovid MEDLINE and Epub Ahead of Print, In-Process & Other Non-Indexed Citations, Daily and Versions, and CINAHL on 25 June 2020. We searched clinical trials databases and the reference lists of retrieved articles for randomised controlled trials (RCTs) and quasi-RCTs.

Selection criteria: 

We included RCTs and quasi-RCTs comparing short (e.g. one or two hours) versus long (e.g. three or four hours) feeding intervals in preterm infants of any birth weight, all or most of whom were less than 32 weeks' gestation. Infants could be of any postnatal age at trial entry, but eligible infants should not have received feeds before study entry, with the exception of minimal enteral feeding. We included studies of nasogastric or orogastric bolus feeding, breast milk or formula, in which the feeding interval is the intervention.

Data collection and analysis: 

We used standard methodological procedures expected by Cochrane. We used the GRADE approach to assess the certainty of evidence. Our primary outcomes were days taken to achieve full enteral feeding and days to regain birth weight. Our other outcomes were duration of hospital stay, episodes of necrotising enterocolitis (NEC) and growth during hospital stay (weight, length and head circumference).

Main results: 

We included four RCTs, involving 417 infants in the review. One study involving 350 infants is awaiting classification. All studies compared two-hourly versus three-hourly feeding interval. The risk of bias of the included studies was generally low, but all studies had high risk of performance bias due to lack of blinding of the intervention.

Three studies were included in meta-analysis for the number of days taken to achieve full enteral feeding (351 participants). The mean days to achieve full feeds was between eight and 11 days. There was little or no difference in days taken to achieve full enteral feeding between two-hourly and three-hourly feeding, but this finding was of low certainty (mean difference (MD) ‒0.62, 95% confidence interval (CI) ‒1.60 to 0.36).

There was low-certainty evidence that the days taken to regain birth weight may be slightly longer in infants receiving two-hourly feeding than in those receiving three-hourly feeding (MD 1.15, 95% CI 0.11 to 2.20; 3 studies, 350 participants).

We are uncertain whether shorter feeding intervals have any effect on any of our secondary outcomes including the duration of hospital stay (MD ‒3.36, 95% CI ‒9.18 to 2.46; 2 studies, 207 participants; very low-certainty evidence) and the risk of NEC (typical risk ratio 1.07, 95% CI 0.54 to 2.11; 4 studies, 417 participants; low-certainty evidence).

No study reported growth during hospital stay.