Any child who needs supplemental nutrition and has functional intestine should be given enteral feeds. This can be by nasogastric, nasoduodenal (post-pyloric) or nasojejunal tube, or by gastrostomy, gastrojejunostomy or jejunostomy tube. Each of these, even the more invasive ones, are preferable to parenteral feeds because of the lower risk of hepatic dysfunction, deep venous thrombosis, and line sepsis. However, surgically placed jejunostomy tubes, especially when done with a roux-en-Y technique, are dangerous due to the risk of volvulus, especially in patients who are neurologically impaired. Feeds are most conveniently given as gastric boluses, which are also said to be more “physiologic.” Nevertheless, children seem to tolerate continuous feedings well and in some situations it clearly more practical. A regimen of bolus feedings during the day and continuous at night is often more convenient for parents and the night feeds allow more calories to be given. The disadvantages of continuous feedings include the need for a pump, the tethering effect of being attached to a tube, and the risk of contamination of the formula with microbes.
Some turn the simple procedure of advancing feeds into a complex art form with convoluted rules and restrictions that border on superstitious (gastric residuals). Enteral feeds are usually advanced very gradually in prematures because of the thought that rapid feeding advancement might precipitate bowel ischemia or NEC. Likewise, infants who have been treated for “medical” NEC are at risk for colonic strictures, which appear to be a risk factor for overwhelming sepsis when feeds are advanced quickly. For most other children, feeds should be advanced as quickly as tolerated – without pain, reflux or diarrhea. Which formula, how much, and by what route should be agreed upon and then feeds started at one quarter to one half of the goal rate. Some prefer to start with a glucose-electrolyte solution or diluted formula but as soon as it is clear that the patient is tolerating even a small amount of these solutions he or she should be switched to the appropriate full-strength formula. Advancing all the way to goal volume with anything other than formula makes little sense, except perhaps in the rare case of the child who is at risk for dehydration and has no intravenous access. (Recall that in infants, the goal rate for hydration is about two thirds of the goal rate for calories.) We usually start with one third volume feeds and then advance to two thirds and then full feeds every 8, 12, or 24 h, depending on how quickly the child is expected to tolerate it. About half of the total volume can usually be given at night as continuous feeds. The key is that the child must be assessed for reflux symptoms, discomfort, severe abdominal distension, and watery diarrhea at every step to be sure that they are tolerating the advancement. It is dangerous to put the schedule on auto-pilot. Gastric residuals and abdominal girths are generally not very accurate in assessing feeding tolerance. As always, it is important to keep the regimen as simple as possible.
Children with intestinal failure (short gut) are at risk for malabsorption and may not tolerate rapid feeding advancement. Similarly, infants with gastroschisis have bowel dysmotility and foreshortened intestine and are notoriously difficult to get up to full feeds except very slowly. It is useful to start with continuous feeds and then gradually consolidate the feeds into boluses after full volume is achieved. In these cases, we start with a very small amount (5 mL/h or less) and then advance by 1 mL/h/day, as long as stool output is less than 15 mL/kg/shift (45 mL/kg/day). More than this and fluid and electrolyte abnormalities become difficult to manage. If there is profuse diarrhea, feeds should be stopped for at least 8 h and then restarted at the last rate that was tolerated for a few days before trying to advance again. Children with proximal high-output stomas can be “re-fed” the effluent through a mucous fistula, in which case the only output that matters is the actual (more distal) stool output.