Influence of gastric residual assessment in preterm neonates on time to achieve enteral feeding (the GRASS trial)—Multi-centre, assessor-blinded randomised clinical trial

Gastric residual measurement, the aspiration and evaluation of stomach contents, is routinely performed prior to feeding preterm infants in the neonatal intensive care unit (NICU) [1, 2]. Despite a lack of evidence confirming benefit, aspirates are used extensively to confirm naso- or orogastric (NG/OG) tube placement, monitor the amount and appearance of previous feed remaining in the stomach and prevent aspiration of stomach content. Residuals are used as an indicator of feeding intolerance. Feed intolerance occurs in preterm infants for a variety of non-pathological reasons including reduced motility and the normal structural and functional immaturity of the gastrointestinal (GI) tract [3‐5]. Historically, large or bile-stained residuals have been seen as an indicatoro of developing necrotising enterocolitis (NEC) [6, 7]. However, there is also potential harm associated with routine checks—loss of necessary gastric enzymes and micronutrients and mucosal irritation and damage [8]. Because of concern about the potential relationship between residuals and GI pathology, their presence may lead to the holding of feeds or delays in advancing volumes, which can in turn delay the removal of central access.

Adequate nutrition in the early neonatal period is a vital factor in the management of very low birth weight (VLBW) infants (birth weight < 1500 g). While nutrition can be provided via parenteral nutrition (PN), extending the requirement for PN, and therefore, intravenous (IV) access can be associated with increased rates of late-onset sepsis (LOS) [9, 10]. Appropriate early extrauterine growth has been associated with fewer medical complications in the neonatal and infant period [11, 12], improved neurodevelopmental outcomes [13, 14] and decreased metabolic complications in later life [15, 16]. Therefore, establishing enteral feeding promptly yet safely is paramount [17].

To date, multiple randomised and observational studies have evaluated the impact of gastric residuals. Although varying results are described, little evidence of benefit has been found. Our aim was to investigate if the exclusion of routine checks of gastric residuals would have an impact on the time taken to achieve enteral feeding of 100 ml/kg/day, in VLBW neonates.

In this trial, we found no statistically significant difference in the establishment of enteral feeds of 100 ml/kg/day of expressed breast milk, by day of life five, in a group of appropriately grown VLBW infants born between 26⁺⁰ and 30⁺⁶ weeks of gestation. There were no differences in prespecified secondary outcomes.

Previous literature in this area has concentrated on outcomes regarding feeding achievement and growth rather than safety outcomes such as mortality, NEC, and LOS, leading to meta-analyses to attempt to answer these questions.

The latest Cochrane review on this topic [23] provided a meta-analysis of studies evaluating the measurement of gastric residuals vs. no residuals. Four studies involving 334 infants were included in this comparison. Three studies included infants < 1500 g, while one allowed infants up to 2 kg. They were unable to do a planned subgroup analysis of VLBW infants due to the data available. They found that routine monitoring of gastric residuals probably has little or no effect on the risk of NEC with moderate certainty of evidence (RR 1.08, 95% CI 0.46 to 2.57). They probably increase time to full feeds, TPN days and risk of invasive infection with little or no difference in all-cause mortality predischarge. They conclude that further trials are warranted to increase certainty and assess longer-term outcomes.

This review included the first randomised trial of residual assessment [8] of 61 infants under 32 weeks and less than 1.25 kg, which found that infants reached feeds of 150 ml/kg/day and had IV access removed almost 6 days earlier if residuals were not checked. This result did not reach statistical significance but was considered potentially clinically significant. Feeding was notably slower than in our cohort, with a mean (SD) time to 120 ml/kg/day of enteral feeding of 16.8 (12.4) days and 14.3 (12.5) days in the control and intervention arms respectively.

This was followed by a similar trial [24] of 143 infants < 1.25 kg and < 32 weeks randomised to gastric residual assessment or no residual assessment, evaluating the effect of checking residuals on weekly enteral nutrition over the first 6 weeks of life. The no-residual group advanced feeding quicker with higher mean weight gain. They were discharged home 8 days earlier with a similar odds ratio for NEC, LOS and death.

While our results are similar to previous trials in failing to show harm if residuals are not checked, although none have been powered for these outcomes, we did not show any statistical or clinical benefit to their avoidance. This may be explained by the difference in feeding practices between units and the evolution of feeding practices over time. In our study, enteral feeding was commencement earlier, with more rapid advancement and earlier achievement of full enteral feeding, consistent with changes in premature feeding guidelines internationally. Quicker establishment of enteral feeds may have meant less impact of residual measurement, being more difficult to reverse enteral feeding to IV in a baby who is fully established on feeds without IV access readily available. Our cohort was a more mature cohort with larger infants, as we had a higher GA and weight cut-off.

Our protocol allowed clinicians to make pragmatic clinical decisions on the management of what they deemed a concerning aspirate and to combine this with the full clinical picture. This highlights the minimal impact routine measurement has on decision-making in NICU. We allowed regular venting of the gastric tube. As most infants were managed on nasal continuous positive airway pressure, this may have decreased episodes of abdominal distension compared to other studies—leading to less diagnosis of feed intolerance and interruption of advancement.

The main concern regarding the removal of routine checks of residuals is the removal of an early indicator of NEC, removing opportunities to intervene early and minimise morbidity. This is based on the results of two older case–control studies on NEC risk factors [6, 7]. The first [6] suggested that a volume over 3 ml may be associated with a higher risk of NEC. A study of 34 infants [7] observed higher volume residuals in the group which later developed NEC (7.5 ml vs. 4 ml). However, a 17-day delay was noted from maximum residual to a diagnosis of NEC, minimising clinical utility.

While considered a benign investigation, gastric residuals contain essential gastric enzymes and micronutrients which are lost if the aspirate is discarded. Hydrochloric acid is essential to the intestinal barrier and necessary to limit bacterial overgrowth. Discarding aspirates may increase inflammation and further alter the preterm gut microbiome, increasing the risk of late-onset sepsis [25, 26]. Refeeding may be done but remains controversial.

The strengths of this study include the randomisation process with stratification, the multi-centre nature and powering to a clinically relevant endpoint. Our findings are limited to the population we studied and may not be generalizable to infants < 26 weeks, with IUGR or critically unwell infants. As with previous studies, we were not powered to detect a difference in NEC.

In the absence of a clinical benefit to routinely monitoring gastric residuals in preterm infants, it may be appropriate to discontinue this practice and instead only monitor residuals when clinical concern of feed intolerance or gastrointestinal pathology arises.