Objective. To test the hypothesis that very low birth weight infants fed by continuous nasogastric gavage (CNG) would achieve full enteral feedings (100 kcal/kg/d) at an earlier postnatal age and have less feeding intolerance (FI) than infants fed by intermittent bolus gavage (IBG).
Methods. Eighty infants were stratified by birth weight (700 to 1000 g and 1001 to 1250 g) and randomized into CNG or IBG feeding groups. CNG infants were comparable with IBG in birth weight, gestational age, sex, race, and day of onset of feeding (5.7 ± 2.1 days vs 5.6 ± 2.2 days, respectively). Feedings were given as undiluted Similac Special Care formula (Ross Laboratories, Columbus, OH) via a specific protocol designed for each 50 to 100 g birth weight category. Feedings were advanced isoenergetically by a maximum of 25 mL/kg/d until an endpoint of 100/kcal/kg/d for at least 48 hours was reached. An infant whose feedings were withheld for >12 hours based on predetermined criteria was considered to have an episode of FI.
Results. Infants in the CNG group reached full enteral feeding at 17.1 ± 8.9 days compared with 15.5 ± 5.5 days in the IBG group; these were not statistically different. Secondary outcome variables such as days to regain birth weight (CNG, 12.6 ± 5 days vs IBG, 12.5 ± 3.7 days), days to reach discharge weight of 2040 g (CNG, 60 ± 13.4 days vs IBG, 62 ± 13.6 days), and number of episodes of FI were not significantly different between feeding methods. FI was primarily associated with birth weight ≤1000 g (71%) vs 1001 to 1250 g (38%).
Conclusion. Feeding methods are associated with similar outcomes when feeding regimens are comparable.
Achievement of adequate nutrition, as soon as possible after birth, is the ultimate goal for all low birth weight infants (<1250 g); however, the best methods by which sufficient enteral nutrition can be provided remain controversial. Continuous nasogastric gavage (CNG) was introduced in 19721 and became widely accepted because of the ease of administration, the potential for sustaining growth,2 and the decreased time for steady weight gain.3 Advantages and disadvantages of different feeding methods have been proposed.4,5 In a survey of neonatologists in 1985, intermittent bolus gavage (IBG) was the preferred method of feeding low birth weight infants.6Prospective controlled studies comparing these two methods are limited in number and either do not focus on very low birth weight infants (<1250 g) or do not include similar isoenergetic feeding protocols.
This study was designed to test the hypothesis that infants who weigh between 700 and 1250 g and are fed by CNG would have improved feeding tolerance and, therefore, would reach full enteral feedings at an earlier postnatal age than those fed by IBG.
The study was conducted between April 1994 and July 1995 in neonates admitted to the Neonatal Intensive Care Unit of Cook County Children's Hospital. The study design and consent forms were approved by the Internal Review Board of Cook County Hospital and informed consent was obtained from either of the parents when the infant was considered eligible for entry into the study.
Patients were eligible for enrollment if their birth weight was between 700 to 1250 g, if they had a soft abdomen and bowel sounds were present, and were considered by the attending neonatologist to be in a hemodynamically stable condition and ready to start enteral nutrition. The infant was included if still on a ventilator but feedings were not started until the umbilical arterial catheter was removed. Exclusion criteria were: Apgar score <3 at 5 minutes, mother had chosen to breastfeed, documented sepsis, necrotizing enterocolitis (NEC), or inability to start feedings before the tenth day of postnatal life. Once enrolled, patients were removed from the study if the route of feeding was changed in error or if the infant required a special formula, developed NEC, was transferred back to the referring hospital, or died before reaching full enteral feedings. All complications were noted even if the feeding data could not be included in the analysis.
The infants were stratified according to birth weight into one of two groups (Group A weighed between 700 and 1000 g and Group B weighed between 1001 and 1250 g). Infants were randomly assigned within each weight group to either CNG or IBG by using sequentially numbered opaque sealed envelopes using a table of random numbers. CNG feedings were delivered via an indwelling 5F polyvinyl nasogastric catheter with a continuous infusion pump (IVAC Medsystem III, San Diego, CA). IBG feedings were given by gravity every 3 hours for 15 to 30 minutes via an indwelling 5F nasogastric catheter similar to that used for CNG. Positioning of the nasogastric tube was not routinely confirmed by radiographs. Catheters in both groups were changed every 3 days per standard nursery protocol. Nonnutritive sucking was continued in both groups of infants to lessen behavioral distress.7 Gastric residual (GR) was measured with the same catheter every 2 hours in the CNG group and every 3 hours (prefeed) in the IBG group as currently practiced in our nursery. All infants were maintained in closed incubators until they weighed approximately 1800 g and the neonatal staff followed standard nursery protocols other than for feeding.
Parenteral nutrition, including lipid emulsion, was started on days 2 to 3 and continued until each infant tolerated full enteral feedings. All infants were fed undiluted 20 cal/oz Similac Special Care formula (Ross Laboratories, Columbus, OH). Feeding protocols were designed for each 50 to 100 g weight category as shown in Table1. Patients were started on 12 to 16 mL/kg/d and the feedings were advanced by a maximum of 25 mL/kg/d. Feedings were advanced at a similar increment of mL/kg/d for both groups. The caloric and protein intake was identical in the two groups. Successful achievement of enteral feedings was defined as the ability to tolerate enteral feedings of 100 kcal/kg/d for at least 48 hours. Nipple feedings were initiated when infants reached a weight of 1500 g; feedings were then advanced by the neonatologist as tolerated by the infant.
Arbitrary guidelines for excessive residual were developed based on consensus among the study staff and the attending neonatologists. Thus, excessive residual was defined in the CNG group as a GR volume >2.5 times the hourly volume of formula when the rate of infusion was <2 mL/h; >1.5 times when the rate was 2 to 3 mL/h; more than the hourly rate when the rate was 3 to 5 mL/h, or more than half when the rate was >5 mL/h. In the IBG group, the residual was defined as excessive when the amount of formula was more than half of the preceding feeding. Except in rare instances, GR was refed in either group.8
Feedings were discontinued for 3 hours if residual was excessive and there were no other clinical findings. Guidelines for withholding feedings for longer periods included two or more of the following: excessive GR, increase in abdominal girth measured at the umbilicus by 2 cm or more9 in 6 hours, guaiac positive stools, dilated loops of bowel, abnormal abdominal roentgenograms, possible sepsis, apnea, and bradycardia occurring more than 3 times per 8 hour shift (cessation of breathing for more than 20 seconds and heart rate <80 beats/min). Any infant whose feedings were withheld for more than 12 hours was considered to have an episode of feeding intolerance (FI). NEC was defined by modified Bell's criteria.10
Daily weight, intake and output, number of stools, number of guaiac positive stools, number of hours that feedings were withheld, episodes of apnea, and bradycardia were recorded. All stools were tested for blood and abdominal circumference was measured every 8 hours routinely and more often depending on clinical status.
Based on the data of Krishnan et al,3 we hypothesized that CNG feedings would result in a 35% decrease in the number of days required to reach full enteral feedings (primary outcome) as compared with IBG feedings. The sample size, based on an α of .05 and a β error of .20, was calculated to be 70 patients. For continuous measures, the two groups were compared using the unpaired Student's t test adjusting for unequal variances if necessary. Categorical measures were compared using the χ2 test or Fisher's exact test when the expected cell sizes were not adequate to meet the test assumptions. All results are reported as mean ± standard deviation.
Ninety-one patients were eligible for the study and 89 were enrolled. Of the 2 patients who were not enrolled, 1 had developed NEC and the other was fed before consent could be obtained. Of the 89 enrolled, 9 (4 CNG and 5 IBG) were not included in the data analysis for the following reasons: of the 4 CNG patients, 2 were diagnosed as having severe congenital syphilis, 1 had been switched to breastfeeding, and 1 required surgery for intestinal malrotation. Of the 5 IBG patients, the protocol had not been followed in 3 patients, 1 patient had been switched to breastfeeding, and 1 was transferred to another hospital before completing the protocol. There were no deaths among the 89 infants and no other adverse outcomes.
Demographic data of the 80 patients (39 CNG, 41 IBG) who completed the study are provided in Table 2. There were 40 patients who weighed ≤1000 g at birth and 40 who weighed between 1001 and 1250 g at birth. No significant differences were seen between the CNG and IBG groups at the onset of the study. Feedings were initiated in the CNG group at 5.7 ± 2.1 days and in the IBG at 5.6 ± 2.2 days.
Infants regained their birth weight (CNG, 12.6 ± 5 days vs IBG, 12.5 ± 3.7 days), reached successful feeding criteria of 100 kcal/kg/d (CNG, 17 ± 8.9 days vs IBG, 15.5 ± 5.5 days), and reached discharge weight of 2040 g (CNG, 60 ± 13.4 days vs IBG, 62 ± 13.6 days) at similar time periods regardless of the feeding methods. There were also no significant differences between CNG and IBG methods when the results were examined within each weight category (Table 3), although the smaller infants took longer than their larger counterparts to reach full enteral feeds (18.7 ± 5.9 vs 12.8 ± 4.6, P< .0001) and discharge weight (68 ± 9.8 vs 48.7 ± 6.6P < .0001). In addition, there were no differences between the groups in the number of days spent on the ventilator (CNG, 8.9 ± 10.9 days vs IBG, 9.4 ± 9.8 days).
FI was diagnosed in 28 patients, 14 in each group. In general, a larger proportion of infants whose feedings were discontinued because of excessive residual were in the CNG group, whereas apnea and bradycardia were more often seen in the IBG group. Of the 14 infants with FI in the CNG group, 4 had suspect NEC (Bell's stage 1) and 2 had Bell's stage 2 or greater; 1 of the 2 required surgery and the other was treated medically. In the IBG group, 1 patient had suspect NEC and 2 had Bell's Stage 2 or greater; 1 of the 2 required surgery and the other was treated medically.
Infants with FI were smaller at birth, and were on the ventilator longer when compared (Table 4) with those without FI. The presence of FI was not a factor in the time to regain birth weight as caloric intake was primarily supplied by parenteral nutrition. FI was associated with time to reach full enteral feedings as well as with time to reach discharge weight most likely because 20 of the 28 infants with FI weighed less than 1000 g at birth.
This study did not confirm our hypothesis that continuous nasogastric feedings would result in a decreased amount of time to achieve full feedings of 100 kcal/kg/d. There were also no significant differences between CNG and IBG methods in days to achieve full feedings nor in time to regain birth weight or to reach discharge weight. These findings are in contrast to those reported in abstract by Krishnan and Satish;3 however, their study was retrospective and not controlled for energy intake. Our results extend the findings of Silvestre et al11 who defined full enteral feedings as 75 kcal/kg/d and discharge weight as 1800 g; in their study there were also no differences between feeding methods in growth and macronutrient retention. Infants were approximately 100 g heavier than in our study and were started on enteral feedings while the umbilical catheter was still in place. Toce et al12used a 20 cal/oz isoenergetic regimen and showed improvement in weight gain in only one subgroup of infants (1000 to 1249 g); there were no differences in infants >1250 g nor in the 10 infants who were <1000 g; all were off the ventilator when feeding protocols were begun. Urrutia and Poole13 reported, in an abstract, similar results between feeding methods in infants <1500 g. Symington et al14 also found no differences between the feeding methods but recommended continuous feeding as less expensive because the catheters in the intermittent group were removed after every feeding. In our study, catheters were replaced every 3 days in both groups of patients; hence cost of catheters was not a serious issue. Bolus feeding on the other hand, was considered superior to continuous feeding by Schanler et al,15 but the study was not controlled and FI was defined solely on the basis of GR.
CNG methods have been recommended as a way of increasing energy efficiency because of improved absorptive capacity by the gut,16,17 as a means of decreasing the amount of time required to reach full feedings3,12 and as the best method for infants with intestinal disease.18 Others have suggested that IBG may be more beneficial because of lower risk of precipitation into the delivery system or because of ease of administration because it does not require as much equipment;4,5 however, this may be compensated by the nursing time required to deliver a bolus feeding. IBG, on the other hand, has been associated with increasing gastroesophageal reflux, abdominal distention, apnea, and bradycardia attributable to increased vagal stimulation and aspiration because of poor gastric motility.4,5,19 Increased cyclic surges of gut hormones are found in infants fed intermittently; however, higher peak levels of insulin, gastrin, and other gut hormones are maintained by continuous feedings.20 The clinical significance of these findings in improving nutrition could not be established as weight gain was not different between intermittent and continuous groups.
In general, the literature suggests that birth weight is usually regained in an inverse relation to birth weight.21 In our study, birth weight was regained at similar times regardless of method of feeding or birth weight category, this was most likely attributable to the early and liberal use of parenteral nutrition (on day 2 or 3) when the infants were considered metabolically stable.22,23
Our hypothesis that infants fed by CNG would achieve caloric feedings earlier than their IBG counterparts was based on our assumption that CNG infants would have less FI. Unlike the findings of Silvestre et al11 the number of infants in our study who developed FI were the same regardless of method of feeding. Although the study sample was not based on the incidence of FI, the sample size was adequate to show some important differences between infants who developed FI and those who did not: infants with FI were more likely to weigh <1000 g and require ventilator support for longer periods of time than their non-FI counterparts. As a result, infants with FI reached full enteral feeds and discharge weights significantly later than those without FI.
Our study may have set higher arbitrary guidelines for excessive GR than used by others. Had these guidelines been lower, an even greater number of infants would have had feedings discontinued for more than 3 hours. On the other hand, none of the infants who had excessive residual without other signs or symptoms developed NEC or FI. Hence, excessive residual, per se, is not an indication for withholding feedings for prolonged periods.
In summary, the method of feeding is not associated with differences in outcome when similar energy intakes are provided and when guidelines for discontinuation of feedings are followed. The final choice of method of feeding remains that of clinical judgment based on the tolerance and the clinical condition of the infant.
- Received June 11, 1996.
- Accepted April 21, 1997.
Reprint requests to (M.K.) Cook County Children's Hospital, Division of Neonatology, 700 South Wood St, Chicago, IL 60612.
Presented in part at the annual meeting of the American Pediatric Society and the Society for Pediatric Research, Washington, DC, May 1996.
- CNG =
- continuous nasogastric gavage •
- IBG =
- intermittent bolus gavage •
- NEC =
- necrotizing enterocolitis •
- GR =
- gastric residual •
- FI =
- feeding intolerance
- ↵Landwirth J. Continuous nasogastric infusion versus total intravenous alimentation. J Pediatr.1972;81:1037–1038. Letter
- Landwirth J
- ↵Krishnan V, Satish M. Continuous (C) vs intermittent (I) nasogastric (N/G) feeding in very low birth weight infants. Pediatr Res.1981;15:537. Abstract
- ↵Kaempf J. Techniques of central feeding in the premature infant. In: Hay WW, ed. Neonatal Nutrition and Metabolism. 9th ed. St Louis, MO; Mosby Year Book; 1991:335–348
- Churella HR,
- Bachhuber WL,
- MacLean WC
- ↵Bell MJ, Ternberg JL, Feigin RD, et al. Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging.Ann Surg. 1978;39:1441. Abstract
- ↵Urrutia JG, Poole E. Continuous nasogastric versus intermittent gavage feedings in very low birth weight infants. Pediatr Res.1983;17:203. Abstract
- ↵Schanler RJ, Shulman RJ, Lau C. Randomized trial of bolus vs continuous tube feeding in premature infants. Pediatr Res.1996;39:1441. Abstract
- ↵Blondheim O, Abbasi S, Fox WW, Bhutani VK. Effect of enteral gavage feeding rate on pulmonary functions of very low birth weight infants.J Pediatr. 1993:751–755
- Haxhija EQ,
- Rosegger H,
- Prechtl HF
- Aynsley-Green A,
- Adrian TE,
- Bloom SR
- Copyright © 1997 American Academy of Pediatrics