Human Milk Versus Formula After Gastroschisis Repair: Effects on Time to Full Feeds and Time to Discharge

ORIGINAL ARTICLE Human milk versus formula after gastroschisis repair: effects on time to full feeds and time to discharge

JA Kohler Sr, AM Perkins and WT Bass

OBJECTIVE: To determine if the choice of enteral feeds after gastroschisis repair relates to the time to achieve full feeds and time to discharge. STUDY DESIGN: A retrospective study of infants with gastroschisis from 2000 to 2010 examined demographics, days at closure, days at initiation of feeds, days to full feeds, time to discharge and length of stay. RESULT: Ninety infants were identified, 22 received (human milk) HM exclusively, 15 were fed 450% HM, 16 were fed o50% HM and 26 were fed only cow milk-based formulas. Infants fed exclusively HM had significantly shorter times to full enteral feedings (median 5 days versus 7 days, P ¼ 0.03). The time from initiation of feedings to hospital discharge, which accounts for initiation age, significantly favored the exclusively HM-fed infants (median 7 days versus 10 days, P ¼ 0.01). CONCLUSION: Exclusive HM feeding after gastroschisis repair decreases time to achieve full enteral feeds and time to discharge.

Journal of Perinatology (2013) 33, 627–630; doi:10.1038/jp.2013.27; published online 21 March 2013 Keywords: total parenteral nutrition; parenteral nutrition-associated liver disease; intestinal permeability

INTRODUCTION METHODS Gastroschisis is a congenital defect of the ventral abdominal wall. Study population The lesion consists of a small full-thickness periumbilical cleft This retrospective study included all patients who were admitted to the usually to the right of the umbilicus, which allows for the neonatal intensive care unit of the Children’s Hospital of The King’s herniation of abdominal contents into the amniotic sac. In addition Daughters during the study period of 1 January 2000 through 31 to small intestines, herniated viscera may include stomach, colon December 2010, who had an International Classification of Disease (ICD)- and the genitourinary organs. The exposed bowel is not covered as 9 procedure code of 54.71 (repair of gastroschisis). This procedure for obtaining gastroschisis patients has been validated previously.10 The in an omphalocele and often becomes matted and thickened with research protocol was approved by the Institutional Review Board of a fibrous peel. Gastroschisis is not associated with known Eastern Virginia Medical School. chromosomal defects, but may be associated with gastrointestinal anomalies such as intestinal atresia, stenosis and malrotation. The incidence of this abnormality is increasing.1–5 As the Study design and data acquisition population affected by this defect continues to grow, the optimal Chart reviews were performed and data were obtained from the attending post-operative feeding regimens for these infants remain neonatologist’s admission history and physical, daily notes and bedside unclear. Much attention has been placed on whether the mode nursing charts. Gestational age was recorded as the number of completed of delivery, type of closure or timing of delivery affects outcomes. gestational weeks. Age at closure, age at initiation of feeds and LOS Following repair, the infant has a prolonged functional ileus were calculated as days of life (DOL), with day of birth serving as DOL 1. Days to full feeds was calculated by subtracting age at initiation of feeds requiring administration of total parenteral nutrition (TPN). from days required to achieve full feeds. Full feeds were defined as an Prolonged TPN has adverse effects on the liver (parenteral enteral feeding volume, either oral or gavage, of 140 ml kg À 1 day À 1 þ / À nutrition-associated liver disease) with elevation of liver enzymes 10 ml kg À 1 day À 1. Days to discharge was calculated by subtracting age at 6 and direct bilirubin. initiation from LOS. Age at initiation was defined as the DOL on which After the return of bowel function, enteral feeds are usually either formula or human milk was first provided for the infant, not by the begun. Previous studies have shown that the earlier the initiation first use of an oral electrolyte solution. Percentage of human milk of enteral feedings, the better the outcome in infants post consumption was calculated from the bedside nursing chart by dividing gastroschisis repair.7,8 Part of this improved outcome can be the total human milk volume by the total enteral feeding volume for the attributable to shorter duration of TPN and reduced length of stay entire hospitalization. (LOS).7 An independent predictor of mortality in these infants is the development of necrotizing enterocolitis. This finding Statistical analysis appears to be mitigated by the timing and type of enteral 9 The data were analyzed using SAS 9.3 software (SAS Institute, Cary, NC, feeds. The focus of this study was to determine if the type of USA). Descriptive statistics (counts, percentages, medians and interquartile enteral feeds, human milk (HM) or formula, improves time to full ranges) were reported for all variables and stratified in two ways both enteral feeds or LOS. based on the percentage of HM consumed. Initially infants were divided

Department of Pediatrics, Children’s Hospital of The King’s Daughters, Eastern Virginia Medical School, Norfolk, VA, USA. Correspondence: Dr JA Kohler Sr, Department of Pediatrics, Children’s Hospital of The King’s Daughters, Eastern Virginia Medical School, 601 Children’s Lane, Norfolk, VA 23507, USA. E-mail: [email protected] Received 25 January 2013; accepted 15 February 2013; published online 21 March 2013 Human milk improves course after gastroschisis JA Kohler Sr et al 628 Table 1. Demographics by percentage of human milk

Percentage of human milk P1 P2 P3 P4 PC

100% 50–99% 1–49% 0%

N 22 15 16 26 Gender, n (%) 0.48a 0.67a 0.43a 0.38a 0.34b Male 10 (45.45) 10 (66.67) 8 (50.00) 11 (42.31) Female 12 (54.55) 5 (33.33) 8 (50.00) 15 (57.69) Mode of delivery, n 0.66a 0.43a 0.21a 0.47a 0.25b (%) Vaginal 13 (59.09) 9 (60.00) 7 (43.75) 12 (46.15) Cesarean section 9 (40.91) 6 (40.00) 9 (56.25) 14 (53.85) Closure type, n (%) 0.86a 0.73a 0.76a 0.69a 0.65b Primary 11 (50.00) 7 (46.67) 6 (37.50) 13 (50.00) Silo 11 (50.00) 8 (53.33) 10 (62.50) 13 (50.00) Birth weight (g), 2670 (2350–2950) 2620 (2270–2950) 2400 (1989–2810) 2463 (2220–2710) 0.10c 0.13d 0.01d 0.09d 0.02e median (IQR) Gestational age 36 (35–38) 37 (35–38) 37 (35–37) 36 (34–37) 0.41c 0.26d 0.15d 0.11d 0.19e (weeks), median (IQR) Age at closure (days), 2 (1–5) 3 (1–8) 5 (2–8) 2 (1–8) 0.58f 0.23g 0.30g 0.84g 0.10h median (IQR) Age at initiation 14 (10–18) 17 (13–20) 17 (13–25) 22 (15–29) 0.003c 0.001d 0.001d 0.01d 0.0001h (days), median (IQR) Abbreviation: IQR, interquartile range. P1 ¼ P-value for test of 100% versus 50–99% versus 1–49% versus 0% human milk; P2 ¼ P-value for test of 100% versus 0–99% human milk; P3 ¼ P-value for test of 50–100% versus 0–49% human milk; P4 ¼ P-value for test of 1–100% versus 0% human milk; PC ¼ P-value for model with a continuous variable for percentage of human milk. aP-value calculated using w2 test. bP-value calculated using logistic regression model. cP-value calculated using analysis of variance. dP-value calculated using two-sample t-test. eP-value calculated using simple linear regression model. fP-value calculated using Kruskal–Wallis test. gP-value calculated using Mann–Whitney U-test. hP-value calculated using Cox proportional hazard model. Bolded values are considered signiﬁcant.

Table 2. Patient outcomes by percentage of human milk consumption

Percentage of human milk P1 P2 P3 P4 PC

100% 50–99% 1–49% 0%

N 22 15 16 26 Days to full feed, median (IQR) 5 (3–9) 7 (5–14) 8 (7–16) 7 (5–13) 0.10a 0.03b 0.08b 0.79b 0.14c LOS (days), median (IQR) 22 (18–24) 26 (22–32) 28 (22–51) 32 (24–54) 0.004a 0.001b 0.001b 0.01b 0.0002c Days to discharge, median (IQR) 7 (5–12) 10 (6–18) 12 (9–25) 10 (8–26) 0.02a 0.01b 0.01b 0.27b 0.01c Abbreviation: IQR, interquartile range. P1 ¼ P-value for test of 100% versus 50–99% versus 1–49% versus 0% human milk; P2 ¼ P-value for test of 100% versus 0–99% human milk; P3 ¼ P-value for test of 50–100% versus 0–49% human milk; P4 ¼ P-value for test of 1–100% versus 0% human milk; PC ¼ P-value for model with a continuous variable for percentage of human milk. aP-value calculated using Kruskal–Wallis test. bP-value calculated using Mann–Whitney U-test. cP-value calculated using Cox proportional hazard model. Bolded values are considered signiﬁcant.

into four groups: 100% HM, 50 to 99% HM, 1 to 49% HM and 0% HM. As these cases exhibit different clinical courses independent of Secondly, infants were divided into only two groups: exclusive HM (100%) enteral feeding, they were also excluded.11,12 and non-exclusive HM (0 to 99%). w2 tests were used to compare proportions for gender, mode of delivery and closure type among (or between) the percentage groups. Analysis of variance, Kruskal–Wallis tests, Demographic comparison of four percentage groups two-sample t-tests, or Mann–Whitney U-tests were used, where appro- The four percentage groups were similar with respect to gender, priate, to assess differences in birth weight, gestational age, age at closure, mode of delivery (vaginal versus cesarean section), type of closure age at initiation, age at full feed, days to full feed, LOS and days to (primary versus silo), gestational age and the age at closure. There discharge. Logistic, linear or Cox proportional hazard regression models was a significant difference in birth weight when comparing assessed for differences in demographics and outcomes, utilizing a continuous variable for percentage of human milk. Finally, analysis of 100% þ 50 to 99% versus 1 to 49% þ 0%, and when compared as a covariance was used to assess the relationship between LOS and continuous variable. There existed a statistically significant percentage of human milk after adjusting for age at initiation or days to difference in age at initiation of enteral feeds between the groups full feed. The level of significance was set at a ¼ 0.05. for all comparisons (Table 1).

Outcome comparison of four percentage groups RESULTS The days to full feed was not significantly different except when Study population comparing 100% human milk to the other percentage groups. A total of 90 infants were identified within the study period. Three The comparison of LOS was significantly different for all compar- infants died prior to the initiation of feeds and were excluded. isons. This mirrored the differences suggested by the differences in Eight infants were classified as complex cases (presence of age at initiation. There was a significant difference in days to intestinal atresias, perforations, necrotic segments or volvulus). discharge between the groups except when comparing any human

Journal of Perinatology (2013), 627 – 630 & 2013 Nature America, Inc. Human milk improves course after gastroschisis JA Kohler Sr et al 629 The post-operative ileus that is seen following gastroschisis repair Table 3. Demographics and outcomes for comparing exclusive requires the concomitant use of TPN. While life-saving, the human milk versus other feeding regimens protracted use of TPN is associated with significant morbidities. Percentage of human milk P2 Sepsis and end-organ dysfunction are significant problems seen in infants on TPN for extended durations.14 Especially concerning is the high risk in developing parenteral nutrition-associated liver 100% 0–99% disease for gastroschisis infants.6 To reduce these risks, TPN duration N 22 57 should be minimized. Our results demonstrated that exclusive Birth weight (g), 2670 (2350–2950) 2540 (2220–2750) 0.13a human milk consumption is associated with faster attainment of full median (IQR) enteral feeds allowing for a decreased duration of TPN. Gestational age 36 (35–38) 36 (34–37) 0.26a Gastroschisis infants are notorious for the unpredictable (weeks), median timing of the return of their bowel function. The time for the (IQR) bowel to ‘wake up’ is unique to each infant and is the hardest b Age at closure 2 (1–5) 4 (1–8) 0.23 factor in their recovery to predict. The in-utero exposure of the (days), median bowel to amniotic fluid results in the bowel damage and (IQR) 15 a decreased motility seen in these infants. One way of Age at initiation 14 (10–18) 18 (13–24) 0.001 (days), median quantifying the recovery of bowel function has been to (IQR) investigate intestinal permeability and time to full enteral 16–18 Age at full feed 19 (17–22) 26 (21–36) 0.002b feeds. Human milk has been shown to decrease intestinal (days), median permeability in preterm infants.19 Human milk feeding has also (IQR) been shown to result in faster attainment of full enteral feeds by Days to full feed, 5 (3–9) 7 (5–14) 0.03b promoting intestinal adaptation.17,18 Our data is consistent with median (IQR) b these findings and supports the idea that human milk is LOS (days), 22 (18–24) 30 (23–39) 0.001 associated with faster attainment of enteral feeds by expediting median (IQR) b the recovery of bowel function. This study represents some of the Days to 7 (5–12) 10 (8–21) 0.01 discharge, first reported data supporting the role of human milk in the median (IQR) recovery of bowel function in infants following repair of gastroschisis. Abbreviation: IQR, interquartile range. As already mentioned, the bowel of infants with gastroschisis P2 ¼ P-value for test of 100% versus 0–99% human milk. aP-value calculated b typically is edematous, inflamed and sometimes covered by a using two-sample t test. P-value calculated using Mann–Whitney U-test. fibrous peel. This renders the intestinal mucosa vulnerable to Bolded values are considered significant. infections such as necrotizing enterocolitis,20 which is six times more likely in exclusively formula-fed infants than exclusively human milk-fed infants and three times more likely than those milk-containing group to exclusive formula consumption (P ¼ 0.27) 9 (Table 2). who received a combination of formula and human milk. Human milk contains components such as pancreatic secretory trypsin inhibitor and transforming growth factor-b2 (TGF-b2), which help Comparison of exclusive human milk (100%) versus non-exclusive provide protection while mediating repair functions and inhibiting human milk (0 to 99%) the mediators of infection.21,22 We suggest that this, along The two groups were similar with respect to birth weight, with the previously discussed decreased intestinal permeability, gestational age and age at closure. There was a significantly is likely the reason that infants with an inflamed intestinal different age at initiation and subsequent age at full feeds mucosa, like that commonly seen with gastroschisis, have a between the two groups. The days to full enteral feeds was decreased incidence of necrotizing enterocolitis when exclusively shorter in the exclusive group (median 5 days) compared with the fed human milk.9 non-exclusive group (median 7 days; P ¼ 0.03). There was a Length of hospital stay has the highest correlation with cost in positive association between the days to full feed and the LOS the care of gastroschisis infants.23 Our data showed that infants (Po0.0001). The LOS was also statistically significant, with the fed exclusively human milk had a significantly shorter LOS versus exclusive group having a median LOS of 22 days and the non- other feeding regimens. Given a median 8-day reduction in LOS, a exclusive group, LOS 30 days (P ¼ 0.001). Analysis of covariance per-diem base neonatal intensive care unit charge of $2645 and 55 demonstrated that exclusive human milk was no longer statisti- infants given formula/partial human milk during the 10-year study cally significantly associated with LOS (P ¼ 0.70) after adjusting for period, this would result in a potential savings to our hospital age at initiation (Po0.0001), so the days to discharge (the days system in excess of one million dollars. While cost should never be between the initiation of enteral feeds and hospital discharge) was the driver of patient care, it should be considered while trying to calculated to adjust for age at initiation. The exclusive group had a determine the best care practices for gastroschisis infants. median days to discharge of 7 days compared with a median of 10 This study is limited because it consisted of a retrospective days in the non-exclusive group (P ¼ 0.01) (Table 3). review. It is also not based on a standard feeding plan outlining standardized rates of enteral feeding advancement. This study also failed to show a graded response to the percent consumption DISCUSSION of human milk when looking at all of the outcome parameters that Gastroschisis is a congenital anomaly with low mortality but we measured. The findings should be validated by prospective significant morbidity.13 These infants require prolonged use of data using standard feeding advancement protocols. TPN and lengthy hospital stays. Operative reduction of the bowel within the abdominal cavity is only the first of many hurdles these babies endure. The prolonged nature of their recovery and the CONCLUSION adverse effects of their treatment highlight areas where Gastroschisis is a complex problem that presents many unresolved improvement in practice can be achieved. Despite advance- issues concerning the care of these infants. The findings of this ments in management, the care of these infants is still associated study support the use of exclusive human milk feedings in post- with significant questions and cost. operative gastroschisis infants. It is recommended that exclusive

& 2013 Nature America, Inc. Journal of Perinatology (2013), 627 – 630 Human milk improves course after gastroschisis JA Kohler Sr et al 630 human milk feeding be encouraged to shorten the duration of 9 Jayanthi S, Seymour P, Puntis JWL, Stringer MD. Necrotizing enterocolitis after TPN, facilitate bowel recovery and decrease the length of hospital gastroschisis repair: a preventable complication? J Pediatr Surg 1998; 33: 705–707. stay. These findings support the inclusion of gastroschisis on the 10 Williams CA, Hauser KW, Correia JA, Frias JL. Ascertainment of gastroschisis using lists of criteria for donor human milk programs. We believe that the ICD-9-CM surgical procedure code. Birth Defects Res A Clin Mol Teratol 2005; the evidence for the benefits of human milk should be used to 73: 646–648. guide clinicians and parents in their choice of enteral feeds. 11 Molik KA, Gingalewski CA, West KW, Rescorla FJ, Scherer III LR, Engum SA et al. Gastroschisis: a plea for risk categorization. J Pediatr Surg 2001; 36: 51–55. 12 Arnold MA, Chang DC, Nabaweesi R, Colombani PM, Bathurst MA, Mon KS et al. Risk stratification of 4344 patients with gastroschisis into simple and complex CONFLICT OF INTEREST categories. J Pediatr Surg 2007; 42: 1520–1525. The authors declare no conflict of interest. 13 Lao OB, Larison C, Garrison MM, Waldhausen JHT, Goldin AB. Outcomes in neonates with gastroschisis in U.S. children’s hospitals. Am J Perinatol 2010; 27: 97–102. ACKNOWLEDGEMENTS 14 Samanta S, Farrer K, Breathnach A, Heath PT. Risk factors for late onset gram- We would like to acknowledge the support and encouragement of this study by the negative infections: a case-control study. Arch Dis Child Fetal Neonatal Ed 2011; late chairman of the Department of Pediatrics, Donald W Lewis, MD. 96: F15–F18. 15 Langer JC, Longaker MT, Crombleholme TM, Bond SJ, Finkbeiner WE, Rudolph CA et al. Etiology of intestinal damage in gastroschisis. I: Effects of amniotic fluid exposure and bowel constriction in a fetal lamb model. J Pediatr Surg 1989; 24: 992–997. DISCLAIMER 16 Shulman RJ, Schanler RJ, Lau C, Heitkemper M, Ou CN, Smith EO. Early feeding, There was no external funding or source of support for this work. antenatal glucocorticoids, and human milk decrease intestinal permeability in preterm infants. Pediatr Res 1998; 44: 519–523. 17 Shulman RJ, Schanler RJ, Lau C, Heitkemper M, Ou CN, Smith EO. Early feeding, REFERENCES feeding tolerance, and lactase activity in preterm infants. J Pediatr 1998; 133: 1 Holland AJA, Walker K, Badawi N. Gastroschisis: an update. Pediatr Surg Int 2010; 645–649. 26: 871–878. 18 Sisk PM, Lovelady CA, Gruber KJ, Dillard RG, O’Shea TM. Human milk consumption 2 Collins SR, Griffin MR, Arbogast PG, Walsh WF, Rush MR, Carter BS et al. The rising and full enteral feeding among infants who weighr1250 grams. Pediatrics 2008; prevalence of gastroschisis and omphalocele in Tennessee. J Pediatr Surg 2007; 121: e1528–e1533. 42: 1221–1224. 19 Taylor SN, Basile LA, Ebeling M, Wagner CL. Intestinal permeability in 3 Vu LT, Nobuhara KK, Laurent C, Shaw GM. Increasing prevalence of gastroschisis: preterm infants by feeding type: mother’s milk versus formula. Breastfeed Med population-based study in California. J Pediatr 2008; 152: 807–811. 2009; 4: 11–15. 4 Laughon M, Meyer R, Bose C, Wall A, Otero E, Heerens A et al. Rising birth 20 Oldham KT, Coran AG, Drongowski RA, Baker PJ, Wesley JR, Polley TZ. The prevalence of gastroschisis. J Perinatol 2003; 23: 291–293. development of necrotizing enterocolitis following repair of gastroschisis: a sur- 5 Alvarez SM, Burd RS. Increasing prevalence of gastroschisis repairs in the United prisingly high incidence. J Pediatr Surg 1988; 23: 945–949. States: 1996-2003. J Pediatr Surg 2007; 42: 943–946. 21 Marchbank T, Weaver G, Nilsen-Hamilton M, Playford RJ. Pancreatic secretory 6 Christensen RD, Henry E, Wiedmeier SE, Burnett J, Lambert DK. Identifying trypsin inhibitor is a major motogenic and protective factor in human breast milk. patients, on the first day of life, at high-risk of developing parenteral nutrition- Am J Physiol Gastrointest Liver Physiol 2009; 296: 697–703. associated liver disease. J Perinatol 2007; 27: 284–290. 22 Rautava S, Nanthakumar NN, Dubert-Ferrandon A, Lu L, Rautava J, Walker WA. 7 Walter-Nicolet E, Rousseau V, Kieffer F, Fusaro F, Bourdaud N, Oucherif S et al. TGF-beta2 induces maturation of immature human intestinal epithelial cells and Neonatal outcome of gastroschisis is mainly influenced by nutritional manage- inhibits inflammatory cytokine responses induced via the NF-kappaB pathway. J ment. J Pediatr Gastr Nutr 2009; 48: 612–617. Pediatr Gastroenterol Nutr 2012; 54: 630–638. 8 Sharp M, Bulsara M, Gollow I, Pemberton P. Gastroschisis: early enteral feeds may 23 Sydorak RM, Nijagal A, Sbragia L, Hirose S, Tsao RH, Phibbs RH et al. Gastroschisis: improve outcomes. J Paediatr Child Health 2000; 36: 472–476. small hole, big cost. J Pediatr Surg 2002; 37: 1669–1672.

Journal of Perinatology (2013), 627 – 630 & 2013 Nature America, Inc.