 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
From the Department of Pediatrics, University of Alabama at Birmingham School of Medicine, Birmingham, Alabama
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1500 g birth weight) infants with and without necrotizing enterocolitis (NEC). Methods. Case-control study compared 51 VLBW infants who had proven NEC (pneumatosis intestinalis, portal venous gas, and/or perforation; excluding spontaneous gastrointestinal perforations) with 102 control subjects (without suspected or proven NEC) who were matched for birth weight, gestational age, race, and sex and were born January 1996 to December 2001. The age in days at diagnosis of NEC was identified in infants with NEC, and feeding characteristics were recorded for the previous 6 days. Feeding characteristics were recorded for control subjects for the corresponding time period.
Results. The median birth weight was 822 g and median gestational age was 26 weeks in both groups. Feeds were started on the fifth day, with a planned increase to full feeds over 10 days (median) in both groups. Median time to full feeds was 13 days in both groups. Median age of onset of NEC was day 24. The total residuals as a percentage of total feed volume (the primary outcome), maximum residual in the previous 6 days, maximum residual as a percentage of the feed, maximum residuals over the 6 days, and the percentage of feeds with residuals were higher in the NEC group. The maximum residual (median [25th–75th centiles]) was as follows: control subjects: 2 mL per feed (0.5–3.5) or 14% of a feed (4–33); NEC group: 4.5 mL per feed (1.5–9.8) or 40% of a feed (24–61). The total residuals as percentage of feeds and the average of maximum residuals increased in the NEC group from the first 3 days to the 3 days before diagnosis of NEC, but a similar increase was not noted for control subjects.
Conclusions. VLBW infants who developed NEC had more gastric residuals. However, there was overlap with the normal control subjects. Of the gastric residual data, the maximum residual seems to be the best predictor for NEC in the subsequent days.
Key Words: feeding intolerance • gastric residuals • necrotizing enterocolitis
Abbreviations: NEC, necrotizing enterocolitis • VLBW, very low birth weight
Necrotizing enterocolitis (NEC) remains a major cause of mortality and morbidity in very low birth weight (VLBW) infants despite recent advances in neonatal care. NEC is of multifactorial pathogenesis and has been associated with enteral feeds,1,2 lack of breastfeeding,3 ischemia,4 and infectious causes.4,5 It is possible that early identification and treatment may reduce unfavorable outcomes of NEC.
Feeding intolerance is extremely common in the premature infant. Its relationship to NEC is poorly understood, although it is often considered as a clinical sign of or precursor to NEC. Feeding intolerance is usually characterized by gastric residuals or aspirates before feeding, emesis, and/or abdominal distention.1,6 However, normal standards for "gastric residuals" have not been established in premature infants, although many empirical guidelines have been stated in various sources.1,7,8 It is not known whether gastric residuals are a benign consequence of delayed gut maturation and motility in this VLBW population or they are associated with subsequent development of NEC. This study was designed to determine whether gastric residuals are elevated in the 6 days before diagnosis of NEC in VLBW infants.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
In our institution during the study period, once NEC was diagnosed, feeds were discontinued and an orogastric tube was placed. Infants were evaluated for sepsis, and broad-spectrum antibiotics were initiated. Frequent abdominal radiographs were taken for the staging of NEC and to observe for intestinal perforation. Frequent arterial blood gas, electrolyte, and blood count monitoring was also done. Pediatric surgical evaluation was obtained in infants with advanced NEC.
All cases were matched with 2 control subjects. The 102 control subjects were matched by gestational age (±1 week), birth weight (±50 g), sex, and race. When several infants were close matches, the 2 infants with the closest birth weight were chosen. Also, infants were matched according to the year of birth to avoid any bias as a result of changes in clinical practice. Once identified from the database, the medical records were reviewed for completeness and to ensure that control subjects did not have any feeding intolerance (feeds withheld for more than 1 day). We ensured that control subjects did not have any feeding intolerance to exclude infants with suspected NEC (stage 1 NEC).
Identical information was obtained on all subjects using a prespecified data collection form in both groups. The day that NEC was diagnosed was recorded as day 0 in the NEC group, and the corresponding age (in postnatal days) was noted. The control subjects for each infant were then evaluated with the corresponding postnatal day being day 0. Data were collected for a total of 7 days. This included day 0 (D0) as described above and the 6 days before D0 (D1, D2, D3, D4, D5, and D6). The gestational age, sex, race, age when NEC diagnosed/corresponding day for control, day of first feeding, and the intended plan of feeding were recorded. Also, the type of formula or breast milk feedings was recorded by classification into 5 categories as: only breast milk (>90%), mostly breast milk (60%–90%), 40% to 60% breast milk, mostly formula (60%–90%), and all formula (>90%).
Data recorded daily included weights, the volume of feeds, the maximum residual amount, residual per feed (as percentage), total volume of residuals, and total amount of residuals per total feed volume per day (as percentage; Table 1). We collected data on daily weights to have the comparisons of weight on D0 (Table 2), and we used the volume of feeds for calculating total residual per total feed volume (Table 3).
|
|
|
Statistics
A sample size of 150 subjects was calculated to provide 80% power at an 
of 0.05 to detect a 20% difference in the total residuals as a percentage of total feed volume (the primary outcome) by the t test, with the assumption that the standard deviation of the mean would be 40% of the mean. The data were analyzed by Sigma Stat 2.0 and graphed on Sigma Plot 4.0. The Mann-Whitney rank sum test and the 
2 test were used for analyses.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Gastric residuals between the 2 groups were compared (Table 3). The total residual as a percentage of total feed volume (the primary outcome) was 1.6% (median; 25–75th centile: 0.75–2.7) in the NEC group and 0.4% (0.08–1.22) in the control group (P < .01). The maximum residual in the 6 days before NEC was 4.5 mL (median; 25–75th centile: 1.5–9.8; mean: 6.4) and 2.0 mL (0.5–3.5; mean: 2.2) in the control group for the same time period (P < .01). The maximum residual as a percentage of the corresponding feed volume was 40% (median; 25–75th centile: 24–61) in the NEC group and 14% (4–33) in the control subjects (Fig 1). The median percentage of feeds with residuals in the NEC group was 12.5% (25–75th centile: 6–21) and 6.3% (2–16) in the control group (P < .01). The average of maximum residuals over the 6 days and the total residuals as a percentage of total feeds were also higher in the NEC group (Table 3). The NEC group demonstrated a median maximum residual of 0.9 mL (25–75th centile: 0.5–2.0) in the initial 3 days and 1.3 mL (0.5–3.0; P < .05) in the last 3 days, whereas the control group maintained a maximum residual amount of 0.5 mL in both the initial and the final 3 days. When comparing the maximum residual amounts in each day in the NEC group and control group, the maximum residual amount was significantly different only on D1 (Fig 2).
|
|
The backward stepwise regression model, which included all infant variables and feeding-related variables, identified birth weight, gestational age, and the maximum residual over the 6 days as the 3 variables that contributed most to NEC. The multiple logistic regression equation developed with these 3 variables was as follows: Logit P = 11.34 + (50 x birth weight in grams) – (0.59 x gestational age in weeks) + (0.33 x maximum residual in milliliters). The odds ratios with confidence intervals for this model were as follows: birth weight: 1.003 (1.000–1.006; P = .07); gestational age: 0.554 (0.376–0.817; P < .01); and maximum residual: 1.386 (1.20–1.60; P < .001).
Fourteen (27.5%) infants died in the NEC group and 5 (4.9%) infants died in the control group (P < .001). Long-term outcomes were not evaluated in this study.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Feeding practices vary with attending physicians in our institution. In general, feedings are initiated in stable infants in the first few days of life. Feeds are generally advanced over 7 to 10 days, depending on the infant’s tolerance. Infants with residuals larger than one third of the feed or with bilious or bloody residuals are evaluated clinically. If there are any concerns, then feeds are withheld for 24 hours. If the abdomen is soft, nontender, and nondistended and normal bowel sounds are heard, then the feed is returned and the infant continues to feed but the feed volume may not be advanced that day.
A limitation of this study is that it was a retrospective study. Therefore, data on other possibly important feeding-related variables (eg, color, consistency of residuals) could not be collected if they were not routinely included on patient charts. In addition, this study was a single-center study. It is possible that variations in patient care practices may lead to different results. This study, however, generates several hypotheses that can be answered by prospective studies.
Gastric residuals are the element of feeding intolerance that can be measured and compared most easily. The magnitude of gastric residual that should be considered abnormal has not been defined adequately to date. Some centers have empirically used gastric residual volumes of more than one third or 30% of feeds.8,12 Malhotra et al13 evaluated mean basal 4-hour gastric residual volumes in 50 healthy preterm infants, 38 appropriate for date and 12 small for date infants with gestational ages of 28 to 36 weeks and observed that there was a marked decrease in residuals from day 4 to day 7. A linear correlation was not found between increases in abdominal girth and gastric residuals.13 When the increase in abdominal girth was at least 2 cm, a gastric residual volume of 23% was observed, and this was considered a warning to withhold feeds.13 Mihatsch et al14 investigated the significance of gastric residuals in the feeding of 99 extremely low birth weight (<1000 g) infants who were fed a standard protocol between days 3 and 14. Gastric residuals up to 2 mL in infants 750 g and up to 3 mL in infants from 750 to 1000 g were tolerated.14
Although larger gastric residual volumes were associated with NEC in our study, it is more likely that larger gastric residuals were an early sign of NEC ("pre-NEC") rather than attributable to a common underlying phenomenon such as immaturity or poor gastric motility, as the increase in gastric residuals was more marked immediately before the diagnosis of NEC. The total residuals as a percentage of total feed volume was increased in the NEC group, but this variable is difficult to use in clinical practice. More commonly, a relationship between the residual volume and the corresponding feed is evaluated. In this study, the maximum residual amount was identified as being most closely associated with NEC. On the basis of the results of this study, a gastric residual volume of <1.5 mL or <25% of a feed (the 25th centile for NEC group) is probably within the range of normal. If, however, the gastric residual volume is >3.5 mL or 33% of a feed (75th centile for control subjects), then this may be associated with a higher risk for NEC.
More gastric residuals were observed in the NEC group despite correction for age, weight, race, and sex. Additional research into the underlying pathophysiology is required to determine the association of gastric residuals with NEC and to evaluate techniques by which NEC can be identified earlier in VLBW infants.
|
FOOTNOTES |
|---|
Reprint requests to (N.A.) 525 New Hillman Bldg, 619 South 19th St, University of Alabama at Birmingham, Birmingham, AL 35249. E-mail: ambal{at}uab.edu
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M Chauhan, G Henderson, and W McGuire Enteral feeding for very low birth weight infants: reducing the risk of necrotising enterocolitis Arch. Dis. Child. Fetal Neonatal Ed., March 1, 2008; 93(2): F162 - F166. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Pietz, B. Achanti, L. Lilien, E. C. Stepka, and S. K. Mehta Prevention of Necrotizing Enterocolitis in Preterm Infants: A 20-Year Experience Pediatrics, January 1, 2007; 119(1): e164 - e170. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Caple, D. Armentrout, V. Huseby, B. Halbardier, J. Garcia, J. W. Sparks, and F. R. Moya Randomized, Controlled Trial of Slow Versus Rapid Feeding Volume Advancement in Preterm Infants Pediatrics, December 1, 2004; 114(6): 1597 - 1600. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. B. Kenton, C. J. Fernandes, C. L. Berseth, B. A. Cobb, W. A. Carlo, and N. Ambalavanan Gastric Residuals in Prediction of Necrotizing Enterocolitis in Very Low Birth Weight Infants Pediatrics, June 1, 2004; 113(6): 1848 - 1849. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2004 American Academy of Pediatrics. All rights reserved. |
||
|
||
