- Email: [email protected]
Outcomes of plastic closure in gastroschisis
Outcomes of plastic closure in gastroschisis Kristine Clodfelter Orion, MD, Michael Krein, BS, Junlin Liao, PhD, Aimen F. Shaaban, MD, Graeme J. Pitcher, MBBCh, and Joel Shilyansky, MD, Iowa City, IA
Background. Gastroschisis is a congenital abdominal wall defect in which the intestines develop outside the abdomen and are exposed to amniotic fluid. When the defect is small, lymphatic, venous, and intestinal obstruction may occur and contribute to the formation of intestinal edema, atresia, ischemia, and a thick inflammatory peel. Treatment requires early coverage of abdominal contents either by primary closure or by the placement of temporary Silastic silo followed by abdominal wall closure. Currently, both traditional suture closure and the sutureless plastic closure are being employed to repair the gastroschisis defect. The goal of the current study is to evaluate plastic closure. We predict no difference will be found in clinical outcomes between plastic closure and traditional suture closure. Methods. A retrospective review of 80 patients treated between 2000 and 2009 was performed. Plastic closure was used in 52 (65%) and traditional suture closure in 28 (35%) babies. The surgical procedure was determined by surgeon preference. Of the 31(39%) babies who required silos, 15 (19%) were treated with plastic closure and 16 (20%) underwent traditional closure. We collected the following demographic data and clinical progression data. Using SAS 9.2 (SAS Institute Inc, Cary, NC), we conducted linear regression, logistic regression, and time to event models to compare the following outcomes: days on ventilator, days to start enteral feeds, days to reach goal enteral feeds, days on total parenteral nutrition, hospital charges, duration of stay, mortality, and complications. Results. The mean duration of follow-up was 11.4 months. Patients spent an average of 6 days on the ventilator. There were 2 mortalities. A multivariate analysis demonstrated that no differences were found between the 2 closures with most of the outcomes; however, when compared with traditional suture closure, those babies treated with plastic closure spent 4 days fewer days on the ventilator (P < .01). Those babies who underwent suture closure were more likely to have an infection or sepsis (odds ratio, 5.15; P < .001). When the entire cohort was considered, no significant difference was found between plastic and suture closure in time to start feeds, time to reach goal feeds, time on parenteral nutrition, hospital charges, duration of stay, or complications. Ventral hernias were noted in 46 (58%) patients, 32 (62%) after plastic closure and 14 (50%) after suture closure (P = .32). Hernia repair was required in 16 (20%) patients, 11 (21%) after plastic closure, and 5 (18%) after traditional repair (P = .32). In the silo cohort, children treated with plastic closure required 7.5(P < .01) fewer days to start enteral feeds than those treated with suture closure. Conclusion. Plastic closure of abdominal wall defects in gastroschisis is effective both as a primary procedure and after silo placement. A multivariate analysis shows plastic closure to be associated with fewer days of mechanical ventilation and less likelihood of developing infection or sepsis. (Surgery 2011;150:177-85.) From the Department of Surgery, Division of Pediatric Surgery, University of Iowa, Iowa City, IA
GASTROSCHISIS IS a congenital defect in the abdominal wall. The worldwide incidence is increasing, especially among young mothers.1,2 In third-world countries, mortality can approach 43% in some studies, with sepsis being the leading cause of Accepted for publication May 11, 2011. Reprint requests: Joel Shilyansky, MD, Division of Pediatric Surgery, University of Iowa Hospitals, 200 Hawkins Drive, Iowa City, IA 52242. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2011 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2011.05.001
death.3 Advances in neonatal intensive care in developed countries have decreased mortality significantly, but even in the United States, children with gastroschisis require lengthy and expensive hospitalizations and still suffer substantial morbidity.4,5 In contrast to omphalocele, the abdominal wall defect in gastroschisis is to the right of the umbilicus. The abdominal viscera in gastroschisis, specifically the intestines, are unprotected and exposed to the amniotic fluid in utero and then to the atmosphere after birth. Not infrequently, the fascial defect is narrow, leading to intestinal, venous, or lymphatic obstruction; bowel edema; atresias; and SURGERY 177
178 Orion et al
the formation of an inflammatory peel. Infectious complications, systemic inflammatory response syndrome, and abdominal compartment syndrome may complicate treatment. Early intestinal coverage and control of infection remain the basic therapeutic tenants guiding the management of children with gastroschisis and have been well established.6 Traditionally, early reduction of intestinal contents into the abdominal cavity and suture closure (SC) of the defect was advocated.7 If reduction was not possible, then a hand-sewn silo would have been used. This approach was accomplished under general anesthesia, as an emergency procedure in the operating room, followed by a period of chemical paralysis and mechanical ventilation. Closure was often complicated by some degree of abdominal compartment syndrome 8,9 Use of the silo became increasingly popular once preformed spring loaded silos became available.10,11 Babies underwent rapid early coverage with a silo, which could be accomplished at the bedside, often without general anesthesia; elective SC of the fascial defect was still performed days later. This approach prolongs time to definitive closure, necessitates general anesthesia, and often requires a trip to the operating room, and still holds the risk of abdominal compartment syndrome.12 Sandler et al13 introduced the plastic closure (PC) in 2004 at our institution. Ten patients were described in the sentinel study. The intestines are reduced into the abdominal domain using gentle manipulation shortly after birth, whereas the child’s ventilation and hemodynamics are carefully followed. A preformed Silastic Silo (Dow Corning, Midland, MI) is placed when intestines cannot be reduced safely. Sequential reduction of the intestines is then performed. The defect is then covered with umbilical cord, which has been purposefully left long at birth, or in its absence, a hydrogel dressing and an occlusive plastic dressing are used. Recent reports have confirmed the feasibility of this sutureless abdominal closure for gastroschisis.14 In the current study, we report the outcomes of 80 babies with gastroschisis treated at a single institution over 9 years. Fifty-two babies, including 31 initially requiring a silo, were treated with PC. MATERIALS AND METHODS Patient management. At the time of delivery or transfer of babies born elsewhere, the eviscerated bowel is covered with moist warm gauze and plastic (saran wrap or bowel bag), and it is supported over the abdomen to minimize vascular occlusion. Intravenous access is established and resuscitation
Surgery August 2011
initiated. Airway control is achieved using endotracheal tube when needed. A nasogastric tube is inserted and placed to low intermittent suction. Ampicillin and gentamycin are both administered and continued for 3 days. A pediatric surgeon evaluated the baby and ascertained whether the abdominal contents can be reduced immediately. If complete reduction is not possible, then a premade Silastic Silo with spring-loaded rings is placed at the bedside. If complete reduction is deemed safe, then the pediatric surgeon chose either a traditional SC or PC. The choice is made based on the surgeon’s personal preference. The patients who underwent SC are taken promptly to the operating room for the procedure. Plastic closure is performed in the neonatal intensive care unit. The umbilical cord is tailored to fit the abdominal wall opening and is used to cover the defect. A Carradres (Medline Industries, Mundelein, IL) hydrogel dressing is used when the cord is not available. Plastic Tegaderm (3M Health Care, St Paul, MN) dressing is used to maintain the intestines in the abdomen. The dressing is changed on the third postprocedure day and usually every third day subsequently. The fascial defect closes spontaneously by circumferential contraction and the skin by secondary intention. When the defect is nearly closed, the dressings are discontinued and a dry dressing is applied. Data collection. We obtained Internal Review Board approval and conducted a retrospective analysis of all patients with ICD 9 coding for abdominal wall defect from 2000 until 2009 and treated at the University of Iowa Children’s Hospital. After excluding children with omphalocele and bladder extrophy, 80 patients with gastroschisis were identified. We collected the following demographics for each of these patients: sex, birth weight, gestational age, prenatal findings on ultrasonography, and age at last follow-up. Patient records were evaluated for the presence of comorbidities that could affect patient outcomes. The following data points were collected: Apgar scores, associated pulmonary, gastrointestinal (GI), cardiac, and other major anomalies, including genetic, neurologic, or genitourinary; presence of infections or sepsis; short gut syndrome; oral aversion; formation of ventral hernia; failure of primary closure requiring a second closure; reintubation; abdominal reoperation, such as hernia repair, exploratory laparotomy, bowel resection; and other nonabdominal operations. We defined infection or sepsis as any positive culture or the need for antibiotic treatment, with the exception of the initial course of gentamycin
Surgery Volume 150, Number 2
and ampicillin for more than 3 days. Abdominal reoperation was any abdominal procedure performed, other than the initial gastroschisis closure, during the hospitalization that immediately followed birth. Finally, we evaluated the following outcomes: days to start enteral feeds, days to reach goal enteral feeds, total days of parenteral nutrition, days on a ventilator, duration of stay, and charges. There were 4 groups of patients: primary SC, primary PC, delayed SC, and delayed PC. Primary closure was achieved on the first day of life, shortly after birth. Delayed closure was performed after initial silo placement and sequential gradual reduction of the intestines. Statistical analysis. Using SAS 9.2, a Chi-square test, Fisher exact test, and Student t test were conducted as appropriate to compare demographics and predisposing factors in univariate analyses. Subsequently, we conducted multiple linear regression, logistic regression, and time to event models to compare the outcomes. In the whole sample models, we included all patients assuming that their fundamental biologic and pathologic processes are the same. We assessed the independent effects of closure type and silo placement by statistically adjusting for covariates. Because there could be fundamental difference between primary and delayed closures, we also conducted subset analysis, where we looked exclusively at patients treated with a silo and compared those who underwent PC with those with SC. Because of limited sample size for the delayed closure subset, we reported any P values less than .10 for the subset analysis. When outcome variables were not distributed normally, we performed log transformation when appropriate. The regression coefficients could be interpreted as percent change or extrapolated into changes caused by a per-unit change in independent variables of an average subject in the sample. We adopted the later approach uniformly for ease of interpretation. Continuous data are expressed as mean ± standard deviation. RESULTS Patients. Eighty babies were included in the analysis, 40 girls and 40 boys (Table I). Of the 80 patients, 52 (65%) underwent PC and 28 (35%) underwent SC (Fig 1). There were 15 boys and 13 girls in the SC group and 25 boys and 27 girls in the PC group. At birth, the babies in the SC group were 2333 g (±583 g) and the babies in the PC group were 2425 g (±563 g). The average gestational age was 35.5 weeks (±16.5 days) and 36.0 weeks (±12.5 days) in the SC and PC groups,
Orion et al 179
respectively. The average Apgar score at 1 min was 6.5 (±2.0) and 5.9 (±2.4) in the SC and PC groups, respectively. In the entire sample, 21 (26%) anomalies were identified at birth (Fig 2). These anomalies were categorized as gastrointestinal, cardiac, and other anomalies (neurologic, musculoskeletal, and urologic). No pulmonary anomalies were noted. There were 14 (27%) anomalies in the PC group and 7 (25%) in the SC group. There were 3 children with more than 1 anomaly. Two of these babies had simultaneous cardiac and GI anomalies, whereas the third had cardiac and neurologic (colpocephaly, agenesis of corpus collosum, and optic nerve hypoplasia) anomalies. There were 8 children with GI anomalies, 3 (6%) in the PC group and 5 (18%) in the SC group. Three of these children had short gut syndrome. One child had gastroschisis with jejunal atresia and a Meckel’s diverticulum. One child with multiple intestinal atresias was found also to have biliary atresia requiring a Kasai procedure. The remaining 3 children were identified as having complex gastroschisis with intestinal atresias requiring resection and/ or ostomy placement. There were 3 children with short gut syndrome. The first was born with closed gastroschisis and eventually succumbed to complications of the disease. The second child was found to have extensive small bowel atresia. The baby had 24 cm viable small bowel starting at the pylorus as well as a microcolon. This baby underwent a successful STEP procedure. He was discharged home on adlib oral feedings on day of life (DOL) 31. The third baby was born with an anteriorly placed anus, jejunal and ileal atresias, cecal necrosis, and a microcolon. The child required multiple surgical procedures, including bowel resections and ostomy revisions resulting in 27 cm functional small bowel. This patient was discharged home on DOL 115 with a combination of enteral feeds and parenteral nutrition. Parenteral and enteral nutrition. We compared the length of time parenteral nutrition was required and the time to reach full enteral feeds in the 2 treatment groups (Table II). Babies in both the SC and PC group required an average of 28 days of parenteral nutrition. After both bivariate and multivariate analysis, there was no statistical difference regarding the duration of time parenteral nutrition was required. Babies in the PC group averaged 21 days, whereas those in the SC group averaged 18 days to start enteral feeds after abdominal closure. Again, no statistical difference was found regarding the days to start enteral feeds.
180 Orion et al
Surgery August 2011
Table I. Demographic and clinical characteristics of the study sample Gastrochisis cohort
Demographics Sex Female Male GSA (days) Birth weight (g) Apgar score 1 min 5 min Comorbidities Anomaly GI Cardiac Lung Other Other abdominal surgeries Other nonabdominal surgeries Short gut syndrome Mortality
Silo subset
Total
Plastic
Suture
P value
Plastic
Suture
P value
40 (50%) 40 (50%) 251 ± 14 2,393 ± 568
27 (52%) 25 (48%) 253 ± 13 2,425 ± 564
13 (46%) 15 (54%) 249 ± 17 2,333 ± 583
.64
6 (38%) 10 (62%) 253 ± 13 2,310 ± 608
.89
.19 .49
6 (40%) 9 (60%) 254 ± 11 2,330 ± 552
6.1 ± 2.4 8.1 ± 1.3
5.9 ± 2. 5 8.0 ± 1.5
6.5 ± 2.1 8.3 ± .7
.26 .47
6.31 ± 2.39 8.38 ± 1.45
6.07 ± 2.15 8.00 ± .65
.78 .36
.77 .92
21 (26%) 8 (10%) 10 (13%) 0 6 (8%) 5 (6%)
14 (27%) 3 (6%) 8 (15%) 0 4 (8%) 3 (6%)
7 (25%) 5 (18%) 2 (7%) 0 2 (7%) 2 (7%)
.85 .12 .48 NA 1.00 1.00
5 (33%) 2 (13%) 3 (20%) 0 0 (0%) 1 (7%)
5 (31%) 3 (19%) 0 (0%) 0 2 (13%) 1 (6%)
.90 1.00 .10 NA .48 1.00
16 (20%)
10 (19%)
6 (21%)
.81
3 (20%)
4 (25%)
1.00
3 (4%)
0 (0%)
3 (11%)
.04
0 (0%)
1 (6%)
1.00
2 (3%)
1 (2%)
1 (4%)
1.00
1 (7%)
0 (0%)
.48
GSA, Gestational age.
Babies in the PC group averaged 38 days, whereas those in the SC group averaged 32 days to reach goal enteral feeds after closure. With or without controlling for covariates, there was no statistical difference with regard to days to reach goal enteral feeds. Significant factors impacting parenteral and enteral nutrition included duration of time mechanical ventilation was required, presence of infection or sepsis, need for second closure, and presence of a cardiac anomaly. Mechanical ventilation. Patients were evaluated for the duration of time mechanical ventilation was required. Babies treated with PC spent an average of 4.9 days, whereas those babies in the SC group spent 8.6 days on the ventilator. After controlling for multiple variables, the difference of 4.0 days fewer on ventilator for patients treated with PC (P < .01). The findings suggest that PC was associated with requirement for a shorter period of mechanical ventilation than SC. Examining other variables that affected the duration of time mechanical ventilation was required, we found patients in the silo cohort spent an average of 5.0 more days on the ventilator. Female sex, gastrointestinal anomalies, second closure, short gut syndrome, and low Apgar score at 1min were also
associated with a significantly prolonged period of time on ventilator. Hospital stay and charges. The duration of hospital stay and cost of care remain high for babies with gastroschisis. The average duration of hospital stay was not statistically different for patients with gastroschisis; the data were 36 (±22) days and 46 (±26) days in the PC and SC groups, respectively. The factors that significantly contributed to greater duration of stay included prematurity, cardiac anomalies, poor oral feeding, and prolonged intubation. In the PC group, the mean charges were $174,000, with a range from $34,000 to $1,681,000. In the SC group, the mean charges were not different at $199,000 with a range from $35,000 to $643,000. After examining other variables, we found on average $105,000 more in charges of patients the silo cohort. Other significant cost drivers include preterm birth ($16,000 per gestation week), gastrointestinal anomalies ($245,000), and need for second closure ($400,000). Mortality. There were two late deaths in the study population. One of the children was the baby with short gut syndrome whose hospital course was complicated by liver failure, which eventually led
Surgery Volume 150, Number 2
Fig 1. Study design: Babies with gastroschisis were stratified into plastic or suture closure. Subset analysis was performed on the babies that were initially treated with a Silo.
to a multi-visceral transplant. He had originally undergone primary SC with and subsequently died at age 22 months. The other child with poor intestinal motility was initially treated with a silo and delayed PC; he died at home in his crib at age 9 months. Autopsy findings were consistent with severe dehydration and hypoperfusion. The low mortality rate in this patient cohort precludes statistical comparison. Morbidity. Complications of gastroschisis or repair included the development of ventral hernia, need for subsequent hernia repair, evidence of infection or sepsis, and need to reopen the abdomen. There were 81 complications in 61 patients (Table III). Sixteen patients had 2 complications, and 2 patients had 3 complications. Ten (38%) children in the SC group and 22 (42%) children in the PC group had documented infection, sepsis, or prolonged requirement for antibiotics. Ventral hernias were noted prior to discharge home in 46 patients (58%), 32 of 52 (62%) after PC. To date, hernia repair was required in only 15 (19%) patients, 11 (21%) after PC and 4 (14%) after SC. Thirty-two of the 80 patients (40%) developed an infection or sepsis. A bivariate analysis did not show a difference in infectious complications between the 2 closure types; after a multivariate analysis, however, the babies in the SC group were 5 times (odds ratio [OR] = 5.15, P < .001) more likely to develop an infection or sepsis than those in the PC group. Three patients, 2 after SC and 1 after PC, required the abdomen to be reopened and a second closure performed at a later time. The indication for opening the abdomen were: increased abdominal pressure following delayed SC in one patient and respiratory failure and SIRS in
Orion et al 181
Fig 2. Associated anomalies: Associated anomalies were found in 21 (26%) of babies with gastroschisis.
one baby treated with primary PC. The third child was born with closed gastroschisis and short-gut syndrome. Given the poor prognosis, the parents initially elected to proceed with palliative care; 1 week later, however, the baby remained stable, and the family elected to proceed with operative intervention, which included a Ladd’s procedure and enteroenterostomy. Silo subset analysis. Thirty-one (39%) babies were treated initially with a silo. Patients treated with a silo required a greater duration of mechanical ventilation (10.64 vs 3.5, P < .001) and accrued greater hospital charges ($249,000 vs $139,000, P < .01). In the PC cohort, 37 (71%) babies were closed primarily, and 15 (29%) required placement of a silo and delayed closure. Sixteen (57%) babies in the SC group were treated with silo and delayed closure. In the silo cohort, PC and SC groups were statistically equivalent when comparing Apgar scores (6.3 ± 2.4 vs 6.2 ± 2.2, P = .78), birth weight (2330 ± 552 vs 2310 ± 608, P = .92), gestational age (254 ± 11 vs 253 ± 13, P = .77), or associated anomalies (5/15 vs 5/16, P = .9). The duration of hospital stay for babies treated initially with silo was 38 ± 14 (days) in the PC group and 50 ± 24 (days) in the SC group (P = .11). Similarly, hospital charges ($266,000 ± $398,000 in the PC group and $232,000 ± $157,000 in the SC group; P = .76), were not statistically different. Babies treated with silo and PC required 8.4 ± 5.4 days of mechanical ventilation, whereas babies treated with silo and SC required 12.9 ± 11.1 days of mechanical ventilation (P = .1; Table IV). Patients required parenteral nutrition for 31 ± 15 days in the PC group and 32 ± 17 days in the SC group (P = .81). Patients started enteral feeds an average of 17 ± 8 days in the PC group and 20 ± 7 days in the SC group (P < .01), which was statistically significant in a multivariate analysis with a 7.5 day advantage with traditional approach.
182 Orion et al
Surgery August 2011
Table II. Multivariate and bivariate analysis of outcomes in babies with gastroschisis Mechanical ventilator (days) SC vs PC (b) Silo vs primary (b) Female GSA (weeks) GI anomaly Cardiac anomaly Second closure Infection Second abdominal surgery Oral aversion Ventilator (days) Reintubation SGS Apgar at 1 min Duration of stay
4.08y 2.08 5.22z 6.64z 2.61* .08 8.32y .94 17.53z 2.40 4.21
9.31* .67
Start enteral feeds (days)x
Parenteral nutrition (days)
Goal enteral feeds (days)x
Infection (OR)
2.67 1.88 1.03 5.36* 3.04
3.08 2.04 4.17 6.13 5.17
2.26 1.56 .84 7.90 7.92
5.15z 2.17 .19* .23y 1.16
1.95 8.72 2.26 3.66
1.57 9.89 33.29* 9.76* 5.30
3.78 10.90 23.78 8.40 5.94
.29
.45* 10.91* 6.32
.53 20.34y 6.07
.37 20.57* 4.64
1.07
Duration of stay (days)
Hospital charges 31,000 ($)x
2.19 7.62 5.7 3.09 0.58 2.95y 9.88 11.77* 9.06 6.32
($35.84) ($6.89) $104.72y $92.63* ($12.25) ($15.67)z $244.88z 48.69 $400.26z ($14.05) ($45.87)
16.34* 1.24y 14.03* 21.79
$109.62 $31.13
.96z
*P < .05. yP < .01. zP < .001. xAnalysis with log transformation for outcomes; effects were calculated as expected change per unit change of independent variables of an average patient in the sample. (b) Bivariate model where only SC vs PC and silo vs primary are controlled. GSA, Gestational age; SGS, short gut syndrome.
Table III. Complications of plastic and SC in gastroschisis patients Complications Any complication Infection/sepsis Hernia Hernia repair Second closure
All 61 32 46 16 3
(76%) (40%) (58%) (20%) (4%)
Plastic 41 22 32 11 1
(79%) (42%) (62%) (21%) (2%)
Although children treated with silo and PC required fewer days to start enteral feeds than those treated with silo and SC, the time to reach goal feeds and discontinue parenteral nutrition was not statistically different. It took an average of 34 ± 14 days in the PC group and 37 ± 15 days in the SC group to reach goal feeds. The average duration in PC and SC groups were 31 ± 15 and 32 ± 17 days, respectively. DISCUSSION The management of gastroschisis remains controversial in part because of long hospital stays and frequent complications associated with this
Suture 20 10 14 5 2
(71%) (36%) (50%) (18%) (7%)
OR [95% confidence interval] .86 1.10 .85 1.15 .51
(.57–1.31) (.80–1.52) (.60–1.19) (.52–2.55) (.21–1.20)
P value .46 .57 .32 .73 .28
anomaly. In the current study, we analyzed the outcomes of babies treated for gastroschisis at a single University Children’s Hospital. The primary goal of the study was to determine the safety and efficacy of PC for abdominal wall defects in gastroschisis. A multivariate analysis was used to compare the PC and SC patient groups. The management approach was based on surgeons’ preference. The patients were candidates for primary closure based on the appearance of the intestines, size of the abdominal wall defect, or hemodynamic status. The patients deemed to be poor candidates for primary closure were treated initially with a preformed, spring-loaded silo and
Orion et al 183
Surgery Volume 150, Number 2
Table IV. Multivariate analysis of outcomes in the silo subset of gastroschisis patients Ventilator (days) Suture vs plastic Female sex GSA Apgar at 1 min GI anomaly Cardiac anomaly Infection Ventilator days Reintubation SGS Second abdominal surgery Second closure Other nonabdominal surgery
7.60* .04 .16 1.06 6.83 3.20 3.16
6.56 18.17* 3.65
Start enteral feeds (days) 7.48z .01 1.21 5.72 .51z 14.11z 21.56z 5.34 9.08
Goal enteral feeds (days)
Parenteral nutrition (days)
2.64 4.17
2.40 7.33
12.30* 10.09 3.95 .67y 17.51z
15.15* 11.77 9.08 .80z 18.49y
16.78z 4.81
16.50y 5.63
*P < .10. yP < .05. zP < .01. SGS, Short gut syndrome.
then underwent either SC or PC once gradual reduction was accomplished. Thirty-one (39%) patients required silo placement as their initial intervention. The findings of this study suggest that both PC and SC might be used to achieve primary abdominal wall closure safely in babies born with gastroschisis. Comparing patients treated with PC and SC using a multivariate statistical analysis, no difference was found between the groups in term of Apgars, gestational age, birth weight, associated anomalies, or comorbidities. The analysis suggests that there was no detectable selection bias between physicians’ preferences of the type of closure, and the 2 treatment groups were comparable. When the outcomes in the entire cohort were analyzed, PC and SC seemed equivalent with regard to duration of hospital stay, time to start enteral feeds, achieve goal enteral feeds, and discontinue parenteral nutrition, as well as hospital charges. The 2 groups differed, however, in the number of days spent on the ventilator and the likelihood of associated infections. The babies treated with PC required lesser duration of mechanical ventilation than babies treated with SC. We speculate that this difference might be related to the lesser intra-abdominal pressures and less impingement on diaphragmatic excursion resulting in better mechanics of breathing in the PC group. The patients treated with PC also had a lower likelihood of having a documented or suspected infection. We speculate that PC allows abdominal drainage, whereas SC is associated
commonly with abdominal wall erythema and systemic inflammatory response. Riboh et al14 reviewed 26 patients treated with PC and analyzed time spent on ventilator, time to initiating enteral feeds, time to discharge from the neonatal intensive care units, and rate of complications. Similar to our findings, the authors demonstrated that sutureless closure was associated with lesser time to extubation. Additional variables that might have contributed to prolonged ventilation in the current study were the presence of GI anomalies, the need for a second abdominal wall closure, low Apgar score at 1 min, and the use of a silo. Interestingly, the presence of cardiac anomalies did not affect the requirement for mechanical ventilation. Prolonged intubation and mechanical ventilation has been shown to increase the risk of tracheal stenosis, ventilator-associated pneumonia, barotraumas, and chronic lung disease in neonates.15,16 Our patients did not sustain complications of prolonged ventilation, but with a mean follow-up of 12 months, we might underestimate the incidence of late adverse events. Weinsheimer et al17 evaluated 99 babies with gastroschisis in the Canadian Pediatric Surgery Network database and identified 9 patients treated using sutureless repair. The authors observed that sutureless closure led to a delay in initiating enteral feeds but interestingly, no increase in the duration of time parenteral nutrition was required and no increase in the duration of stay. We found that the time to initiating enteral feeds was equivalent in our analysis. The difference between the studies might stem from the difference in the
184 Orion et al
approach to initiating enteral feeds, the statistical models used for analysis, and the variables tested. All babies treated with PC, by definition, had a ventral hernia initially, but only 60% had ventral hernias noted either at discharge from the hospital or at their follow-up clinic visit. Surprisingly, only 1 in 5 babies required repair, which was the same as the rate of ventral hernia repair in the SC group. Similar findings were noted for babies treated with silo followed by PC. It seems that the gastroschisis fascial defect contracts and often closes spontaneously in a manner similar to simple pediatric umbilical hernias. Even when ventral herniorrhaphy is required, it could be performed in an elective outpatient setting with minimal morbidity when contrasted with suture repair of gastroschisis in the newborn period. This course in turn might also decrease the cost of the initial hospitalization. The use of preformed silos and the staged reduction of gastroschisis followed by SC have been reported commonly.10,18 Previous studies comparing primary SC with silo with delayed SC suggested that babies in the silo group required a greater duration of mechanical ventilation, hospital stay, and requirement for parenteral nutrition as well as greater hospital charges.12,19,20 Preformed silos can distend the abdominal wall defect, at times requiring a tight closure or the use of prosthetic material to achieve coverage. PC after the placement of a preformed silo has not been reported to date.14 In this study, 15 babies underwent PC after silo placement successfully. The fascial defect in babies treated with PC closed rapidly without delaying discharge from hospital. We compared the outcome of babies treated with silo and PC with babies treated with silo and SC. The traditional staged reduction of gastroschisis requires 2 procedures with the definitive SC performed in the operating room. Transporting ill neonates is time and resource intensive as well as potentially risky. Mobilizing the operating room adds additional drag on constrained hospital and personnel resources. In contrast, silo placement followed by PC, could be performed in the neonatal intensive care unit without a single trip to the operating room.19 The statistical power in the analysis of the silo subset was not sufficient to define the differences in outcome between PC and SC groups because of the relatively small sample size. Babies undergoing PC after silo placement, however, were noted to spend 7.6 fewer days on the ventilator than those undergoing SC. This difference is consistent with our findings for the entire sample of gastroschisis patients. We used a P = .10 as the cutoff for statistical significance because of concern over
Surgery August 2011
sample size, which makes our statement less conclusive. Children treated with silo and PC also required fewer days to start enteral feeds compared with silo followed by SC. Earlier initiation of enteral feeds could decrease the risk of cholestatic liver disease, bacterial overgrowth, and sepsis. No statistical differences were found in the duration of stay, requirement for parenteral nutrition, or time to achieve full enteral feeds, suggesting that gastroschisis associated gastrointestinal dysmotility has greater impact on these outcomes. The treatment of babies with gastroschisis has evolved to a point where long-term survival is expected for most babies. However, a high morbidity is still associated with this anomaly. In our series, 76% of children suffered at least 1 complication. The hospital course is long and expensive with a prolonged period of mechanical ventilation and parenteral nutrition. Our analysis suggests that PC can be a definitive and safe method for treating gastroschisis in babies. PC could contribute to the safety of primary closure and lead to increased rate of primary repair. The procedure can usually be performed in the neonatal intensive care unit in minutes without the need to mobilize the transport team, anesthesia, or the operating room staff. Fewer staged repairs and emergency procedures in the operating room might reduce resource use. In developing countries, where neonatologists, neonatal intensive care units, pediatric surgeons, pediatric anesthesiologists, and mechanical ventilators may not be readily available, PC or silo placement followed by PC might represent a potential approach to decrease the cost as well as the morbidity and mortality of gastroschisis.
REFERENCES 1. Lund SC, Bauer K, Berrios M. Gastroschisis. Incidence, complications, and clinical management in the neonatal intensive care unit. J Perinat Neonat Nurs 2007;2:63-8. 2. Loane M, Dolk H, Bradbury I. EUROCAT working group. Increasing prevalence of gastroschisis in Europe 19802002: a phenomenon restricted to young mothers? Paediatr Perinat Epidemiol 2007;21:363-9. 3. Sekabira J, Hadley GP. Gastroschisis: a third world perspective. Pediatr Surg Int 2009;25:327-9. 4. Sydorak RM, Nijagal A, Sbragia L, Hirose S, Tsao K, Phibbs RH, et al. Gastroschisis: small hole, big cost. J Pediatr Surg 2002;37:1669-72. 5. 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;27:97-101. 6. Howell KK. Understanding gastroschisis: an abdominal wall defect. Neonatal Netw 1998;17:17-25. 7. Canty TG, Collins DL. Primary fascial closure in infants with gastroschisis and omphalocele: a superior approach. J Pediatr Surg 1983;18:707-12.
Surgery Volume 150, Number 2
8. Lacey SR, Carris LA, Beyer AJ III, Azizkhan RG. Bladder pressure monitoring significantly enhances care of infants with abdominal wall defects: a prospective clinical study. J Pediatr Sug 1993;28:1370-4. 9. Dimitriou G, Greenough A, Giffin F, Davenport M, Nicolaides KH. Temporary impairment of lung function in infants with anterior abdominal wall defects who have undergone surgery. J Pediatr Surg 1996;31:670-2. 10. Fischer JD, Chun K, Moores DC, Andrews HG. Gastroshisis: a simple technique for staged silo closure. J Pediatr Surg 1995;30:1169-71. 11. Pastor AC, Phillips JD, Fenton SJ, Meyers RL, Lamm AW, Raval MV, et al. Routine use of a SILASTIC spring-loaded silo for infants with gastroschisis: a multicenter randomized controlled trial. J Pediatr Surg 2008;43:1807-12. 12. Lobo JD, Kim AC, Davis RP, Segura BJ, Alpert H, Teitelbaum DH, et al. No free ride? The hidden costs of delayed operative management using spring-loaded silo for gastroschisis. J Pediatr Surg 2010;45:1426-32. 13. Sandler A, Lawrence J, Meehan J, Phearman L, Soper R. A ‘‘plastic’’ sutureless abdominal wall closure in gastroschisis. J Pediatr Surg 2004;39:738-41.
Orion et al 185
14. Riboh J, Abrajano CT, Garber K, Hartman G, Butler MA, Albanese CT, et al. Outcomes of sutureless gastroschisis closure. J Pediatr Surg 2009;44:1947-51. 15. Srinivasan R, Asselin J, Gildengorin G, Wiener-Kronish J, Flori HR. A prospective study of ventilator-associated pneumonia in children. Pediatrics 2009;123:1108-15. 16. Ho AS, Koltai PJ. Pediatric tracheal stenosis. Otolaryngol Clin North Am 2008;41:999-1021. 17. Weinsheimer RL, Yanchar NL, Bouchard SB, Kim PK, Laberge J, Skarsgard ED, et al. Gastroschisis closure- does method really matter? J Pediatr Surg 2008;43:874-8. 18. Lansdale N, Hill R, Gull-Zamir S, Drewett M, Parkinson E, Davenport M, et al. Staged reduction of gastroschisis using preformed silos: practicalities and problems. J Pediatr Surg 2009;44:2126-9. 19. McNamara WF, Hartin CW, Escobar MA, Lee Y. Outcome differences between gastroschisis repair methods. J Surg Res 2011;165:19-24. 20. Bayard D, Ramones T, Phillips SE, Leys CM, Rauth T, Yang EY. Method to our madness: an 18-year retrospective analysis on gastroschisis closure. J Pediatr Surg 2010;45: 579-84.
Background. Gastroschisis is a congenital abdominal wall defect in which the intestines develop outside the abdomen and are exposed to amniotic fluid. When the defect is small, lymphatic, venous, and intestinal obstruction may occur and contribute to the formation of intestinal edema, atresia, ischemia, and a thick inflammatory peel. Treatment requires early coverage of abdominal contents either by primary closure or by the placement of temporary Silastic silo followed by abdominal wall closure. Currently, both traditional suture closure and the sutureless plastic closure are being employed to repair the gastroschisis defect. The goal of the current study is to evaluate plastic closure. We predict no difference will be found in clinical outcomes between plastic closure and traditional suture closure. Methods. A retrospective review of 80 patients treated between 2000 and 2009 was performed. Plastic closure was used in 52 (65%) and traditional suture closure in 28 (35%) babies. The surgical procedure was determined by surgeon preference. Of the 31(39%) babies who required silos, 15 (19%) were treated with plastic closure and 16 (20%) underwent traditional closure. We collected the following demographic data and clinical progression data. Using SAS 9.2 (SAS Institute Inc, Cary, NC), we conducted linear regression, logistic regression, and time to event models to compare the following outcomes: days on ventilator, days to start enteral feeds, days to reach goal enteral feeds, days on total parenteral nutrition, hospital charges, duration of stay, mortality, and complications. Results. The mean duration of follow-up was 11.4 months. Patients spent an average of 6 days on the ventilator. There were 2 mortalities. A multivariate analysis demonstrated that no differences were found between the 2 closures with most of the outcomes; however, when compared with traditional suture closure, those babies treated with plastic closure spent 4 days fewer days on the ventilator (P < .01). Those babies who underwent suture closure were more likely to have an infection or sepsis (odds ratio, 5.15; P < .001). When the entire cohort was considered, no significant difference was found between plastic and suture closure in time to start feeds, time to reach goal feeds, time on parenteral nutrition, hospital charges, duration of stay, or complications. Ventral hernias were noted in 46 (58%) patients, 32 (62%) after plastic closure and 14 (50%) after suture closure (P = .32). Hernia repair was required in 16 (20%) patients, 11 (21%) after plastic closure, and 5 (18%) after traditional repair (P = .32). In the silo cohort, children treated with plastic closure required 7.5(P < .01) fewer days to start enteral feeds than those treated with suture closure. Conclusion. Plastic closure of abdominal wall defects in gastroschisis is effective both as a primary procedure and after silo placement. A multivariate analysis shows plastic closure to be associated with fewer days of mechanical ventilation and less likelihood of developing infection or sepsis. (Surgery 2011;150:177-85.) From the Department of Surgery, Division of Pediatric Surgery, University of Iowa, Iowa City, IA
GASTROSCHISIS IS a congenital defect in the abdominal wall. The worldwide incidence is increasing, especially among young mothers.1,2 In third-world countries, mortality can approach 43% in some studies, with sepsis being the leading cause of Accepted for publication May 11, 2011. Reprint requests: Joel Shilyansky, MD, Division of Pediatric Surgery, University of Iowa Hospitals, 200 Hawkins Drive, Iowa City, IA 52242. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2011 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2011.05.001
death.3 Advances in neonatal intensive care in developed countries have decreased mortality significantly, but even in the United States, children with gastroschisis require lengthy and expensive hospitalizations and still suffer substantial morbidity.4,5 In contrast to omphalocele, the abdominal wall defect in gastroschisis is to the right of the umbilicus. The abdominal viscera in gastroschisis, specifically the intestines, are unprotected and exposed to the amniotic fluid in utero and then to the atmosphere after birth. Not infrequently, the fascial defect is narrow, leading to intestinal, venous, or lymphatic obstruction; bowel edema; atresias; and SURGERY 177
178 Orion et al
the formation of an inflammatory peel. Infectious complications, systemic inflammatory response syndrome, and abdominal compartment syndrome may complicate treatment. Early intestinal coverage and control of infection remain the basic therapeutic tenants guiding the management of children with gastroschisis and have been well established.6 Traditionally, early reduction of intestinal contents into the abdominal cavity and suture closure (SC) of the defect was advocated.7 If reduction was not possible, then a hand-sewn silo would have been used. This approach was accomplished under general anesthesia, as an emergency procedure in the operating room, followed by a period of chemical paralysis and mechanical ventilation. Closure was often complicated by some degree of abdominal compartment syndrome 8,9 Use of the silo became increasingly popular once preformed spring loaded silos became available.10,11 Babies underwent rapid early coverage with a silo, which could be accomplished at the bedside, often without general anesthesia; elective SC of the fascial defect was still performed days later. This approach prolongs time to definitive closure, necessitates general anesthesia, and often requires a trip to the operating room, and still holds the risk of abdominal compartment syndrome.12 Sandler et al13 introduced the plastic closure (PC) in 2004 at our institution. Ten patients were described in the sentinel study. The intestines are reduced into the abdominal domain using gentle manipulation shortly after birth, whereas the child’s ventilation and hemodynamics are carefully followed. A preformed Silastic Silo (Dow Corning, Midland, MI) is placed when intestines cannot be reduced safely. Sequential reduction of the intestines is then performed. The defect is then covered with umbilical cord, which has been purposefully left long at birth, or in its absence, a hydrogel dressing and an occlusive plastic dressing are used. Recent reports have confirmed the feasibility of this sutureless abdominal closure for gastroschisis.14 In the current study, we report the outcomes of 80 babies with gastroschisis treated at a single institution over 9 years. Fifty-two babies, including 31 initially requiring a silo, were treated with PC. MATERIALS AND METHODS Patient management. At the time of delivery or transfer of babies born elsewhere, the eviscerated bowel is covered with moist warm gauze and plastic (saran wrap or bowel bag), and it is supported over the abdomen to minimize vascular occlusion. Intravenous access is established and resuscitation
Surgery August 2011
initiated. Airway control is achieved using endotracheal tube when needed. A nasogastric tube is inserted and placed to low intermittent suction. Ampicillin and gentamycin are both administered and continued for 3 days. A pediatric surgeon evaluated the baby and ascertained whether the abdominal contents can be reduced immediately. If complete reduction is not possible, then a premade Silastic Silo with spring-loaded rings is placed at the bedside. If complete reduction is deemed safe, then the pediatric surgeon chose either a traditional SC or PC. The choice is made based on the surgeon’s personal preference. The patients who underwent SC are taken promptly to the operating room for the procedure. Plastic closure is performed in the neonatal intensive care unit. The umbilical cord is tailored to fit the abdominal wall opening and is used to cover the defect. A Carradres (Medline Industries, Mundelein, IL) hydrogel dressing is used when the cord is not available. Plastic Tegaderm (3M Health Care, St Paul, MN) dressing is used to maintain the intestines in the abdomen. The dressing is changed on the third postprocedure day and usually every third day subsequently. The fascial defect closes spontaneously by circumferential contraction and the skin by secondary intention. When the defect is nearly closed, the dressings are discontinued and a dry dressing is applied. Data collection. We obtained Internal Review Board approval and conducted a retrospective analysis of all patients with ICD 9 coding for abdominal wall defect from 2000 until 2009 and treated at the University of Iowa Children’s Hospital. After excluding children with omphalocele and bladder extrophy, 80 patients with gastroschisis were identified. We collected the following demographics for each of these patients: sex, birth weight, gestational age, prenatal findings on ultrasonography, and age at last follow-up. Patient records were evaluated for the presence of comorbidities that could affect patient outcomes. The following data points were collected: Apgar scores, associated pulmonary, gastrointestinal (GI), cardiac, and other major anomalies, including genetic, neurologic, or genitourinary; presence of infections or sepsis; short gut syndrome; oral aversion; formation of ventral hernia; failure of primary closure requiring a second closure; reintubation; abdominal reoperation, such as hernia repair, exploratory laparotomy, bowel resection; and other nonabdominal operations. We defined infection or sepsis as any positive culture or the need for antibiotic treatment, with the exception of the initial course of gentamycin
Surgery Volume 150, Number 2
and ampicillin for more than 3 days. Abdominal reoperation was any abdominal procedure performed, other than the initial gastroschisis closure, during the hospitalization that immediately followed birth. Finally, we evaluated the following outcomes: days to start enteral feeds, days to reach goal enteral feeds, total days of parenteral nutrition, days on a ventilator, duration of stay, and charges. There were 4 groups of patients: primary SC, primary PC, delayed SC, and delayed PC. Primary closure was achieved on the first day of life, shortly after birth. Delayed closure was performed after initial silo placement and sequential gradual reduction of the intestines. Statistical analysis. Using SAS 9.2, a Chi-square test, Fisher exact test, and Student t test were conducted as appropriate to compare demographics and predisposing factors in univariate analyses. Subsequently, we conducted multiple linear regression, logistic regression, and time to event models to compare the outcomes. In the whole sample models, we included all patients assuming that their fundamental biologic and pathologic processes are the same. We assessed the independent effects of closure type and silo placement by statistically adjusting for covariates. Because there could be fundamental difference between primary and delayed closures, we also conducted subset analysis, where we looked exclusively at patients treated with a silo and compared those who underwent PC with those with SC. Because of limited sample size for the delayed closure subset, we reported any P values less than .10 for the subset analysis. When outcome variables were not distributed normally, we performed log transformation when appropriate. The regression coefficients could be interpreted as percent change or extrapolated into changes caused by a per-unit change in independent variables of an average subject in the sample. We adopted the later approach uniformly for ease of interpretation. Continuous data are expressed as mean ± standard deviation. RESULTS Patients. Eighty babies were included in the analysis, 40 girls and 40 boys (Table I). Of the 80 patients, 52 (65%) underwent PC and 28 (35%) underwent SC (Fig 1). There were 15 boys and 13 girls in the SC group and 25 boys and 27 girls in the PC group. At birth, the babies in the SC group were 2333 g (±583 g) and the babies in the PC group were 2425 g (±563 g). The average gestational age was 35.5 weeks (±16.5 days) and 36.0 weeks (±12.5 days) in the SC and PC groups,
Orion et al 179
respectively. The average Apgar score at 1 min was 6.5 (±2.0) and 5.9 (±2.4) in the SC and PC groups, respectively. In the entire sample, 21 (26%) anomalies were identified at birth (Fig 2). These anomalies were categorized as gastrointestinal, cardiac, and other anomalies (neurologic, musculoskeletal, and urologic). No pulmonary anomalies were noted. There were 14 (27%) anomalies in the PC group and 7 (25%) in the SC group. There were 3 children with more than 1 anomaly. Two of these babies had simultaneous cardiac and GI anomalies, whereas the third had cardiac and neurologic (colpocephaly, agenesis of corpus collosum, and optic nerve hypoplasia) anomalies. There were 8 children with GI anomalies, 3 (6%) in the PC group and 5 (18%) in the SC group. Three of these children had short gut syndrome. One child had gastroschisis with jejunal atresia and a Meckel’s diverticulum. One child with multiple intestinal atresias was found also to have biliary atresia requiring a Kasai procedure. The remaining 3 children were identified as having complex gastroschisis with intestinal atresias requiring resection and/ or ostomy placement. There were 3 children with short gut syndrome. The first was born with closed gastroschisis and eventually succumbed to complications of the disease. The second child was found to have extensive small bowel atresia. The baby had 24 cm viable small bowel starting at the pylorus as well as a microcolon. This baby underwent a successful STEP procedure. He was discharged home on adlib oral feedings on day of life (DOL) 31. The third baby was born with an anteriorly placed anus, jejunal and ileal atresias, cecal necrosis, and a microcolon. The child required multiple surgical procedures, including bowel resections and ostomy revisions resulting in 27 cm functional small bowel. This patient was discharged home on DOL 115 with a combination of enteral feeds and parenteral nutrition. Parenteral and enteral nutrition. We compared the length of time parenteral nutrition was required and the time to reach full enteral feeds in the 2 treatment groups (Table II). Babies in both the SC and PC group required an average of 28 days of parenteral nutrition. After both bivariate and multivariate analysis, there was no statistical difference regarding the duration of time parenteral nutrition was required. Babies in the PC group averaged 21 days, whereas those in the SC group averaged 18 days to start enteral feeds after abdominal closure. Again, no statistical difference was found regarding the days to start enteral feeds.
180 Orion et al
Surgery August 2011
Table I. Demographic and clinical characteristics of the study sample Gastrochisis cohort
Demographics Sex Female Male GSA (days) Birth weight (g) Apgar score 1 min 5 min Comorbidities Anomaly GI Cardiac Lung Other Other abdominal surgeries Other nonabdominal surgeries Short gut syndrome Mortality
Silo subset
Total
Plastic
Suture
P value
Plastic
Suture
P value
40 (50%) 40 (50%) 251 ± 14 2,393 ± 568
27 (52%) 25 (48%) 253 ± 13 2,425 ± 564
13 (46%) 15 (54%) 249 ± 17 2,333 ± 583
.64
6 (38%) 10 (62%) 253 ± 13 2,310 ± 608
.89
.19 .49
6 (40%) 9 (60%) 254 ± 11 2,330 ± 552
6.1 ± 2.4 8.1 ± 1.3
5.9 ± 2. 5 8.0 ± 1.5
6.5 ± 2.1 8.3 ± .7
.26 .47
6.31 ± 2.39 8.38 ± 1.45
6.07 ± 2.15 8.00 ± .65
.78 .36
.77 .92
21 (26%) 8 (10%) 10 (13%) 0 6 (8%) 5 (6%)
14 (27%) 3 (6%) 8 (15%) 0 4 (8%) 3 (6%)
7 (25%) 5 (18%) 2 (7%) 0 2 (7%) 2 (7%)
.85 .12 .48 NA 1.00 1.00
5 (33%) 2 (13%) 3 (20%) 0 0 (0%) 1 (7%)
5 (31%) 3 (19%) 0 (0%) 0 2 (13%) 1 (6%)
.90 1.00 .10 NA .48 1.00
16 (20%)
10 (19%)
6 (21%)
.81
3 (20%)
4 (25%)
1.00
3 (4%)
0 (0%)
3 (11%)
.04
0 (0%)
1 (6%)
1.00
2 (3%)
1 (2%)
1 (4%)
1.00
1 (7%)
0 (0%)
.48
GSA, Gestational age.
Babies in the PC group averaged 38 days, whereas those in the SC group averaged 32 days to reach goal enteral feeds after closure. With or without controlling for covariates, there was no statistical difference with regard to days to reach goal enteral feeds. Significant factors impacting parenteral and enteral nutrition included duration of time mechanical ventilation was required, presence of infection or sepsis, need for second closure, and presence of a cardiac anomaly. Mechanical ventilation. Patients were evaluated for the duration of time mechanical ventilation was required. Babies treated with PC spent an average of 4.9 days, whereas those babies in the SC group spent 8.6 days on the ventilator. After controlling for multiple variables, the difference of 4.0 days fewer on ventilator for patients treated with PC (P < .01). The findings suggest that PC was associated with requirement for a shorter period of mechanical ventilation than SC. Examining other variables that affected the duration of time mechanical ventilation was required, we found patients in the silo cohort spent an average of 5.0 more days on the ventilator. Female sex, gastrointestinal anomalies, second closure, short gut syndrome, and low Apgar score at 1min were also
associated with a significantly prolonged period of time on ventilator. Hospital stay and charges. The duration of hospital stay and cost of care remain high for babies with gastroschisis. The average duration of hospital stay was not statistically different for patients with gastroschisis; the data were 36 (±22) days and 46 (±26) days in the PC and SC groups, respectively. The factors that significantly contributed to greater duration of stay included prematurity, cardiac anomalies, poor oral feeding, and prolonged intubation. In the PC group, the mean charges were $174,000, with a range from $34,000 to $1,681,000. In the SC group, the mean charges were not different at $199,000 with a range from $35,000 to $643,000. After examining other variables, we found on average $105,000 more in charges of patients the silo cohort. Other significant cost drivers include preterm birth ($16,000 per gestation week), gastrointestinal anomalies ($245,000), and need for second closure ($400,000). Mortality. There were two late deaths in the study population. One of the children was the baby with short gut syndrome whose hospital course was complicated by liver failure, which eventually led
Surgery Volume 150, Number 2
Fig 1. Study design: Babies with gastroschisis were stratified into plastic or suture closure. Subset analysis was performed on the babies that were initially treated with a Silo.
to a multi-visceral transplant. He had originally undergone primary SC with and subsequently died at age 22 months. The other child with poor intestinal motility was initially treated with a silo and delayed PC; he died at home in his crib at age 9 months. Autopsy findings were consistent with severe dehydration and hypoperfusion. The low mortality rate in this patient cohort precludes statistical comparison. Morbidity. Complications of gastroschisis or repair included the development of ventral hernia, need for subsequent hernia repair, evidence of infection or sepsis, and need to reopen the abdomen. There were 81 complications in 61 patients (Table III). Sixteen patients had 2 complications, and 2 patients had 3 complications. Ten (38%) children in the SC group and 22 (42%) children in the PC group had documented infection, sepsis, or prolonged requirement for antibiotics. Ventral hernias were noted prior to discharge home in 46 patients (58%), 32 of 52 (62%) after PC. To date, hernia repair was required in only 15 (19%) patients, 11 (21%) after PC and 4 (14%) after SC. Thirty-two of the 80 patients (40%) developed an infection or sepsis. A bivariate analysis did not show a difference in infectious complications between the 2 closure types; after a multivariate analysis, however, the babies in the SC group were 5 times (odds ratio [OR] = 5.15, P < .001) more likely to develop an infection or sepsis than those in the PC group. Three patients, 2 after SC and 1 after PC, required the abdomen to be reopened and a second closure performed at a later time. The indication for opening the abdomen were: increased abdominal pressure following delayed SC in one patient and respiratory failure and SIRS in
Orion et al 181
Fig 2. Associated anomalies: Associated anomalies were found in 21 (26%) of babies with gastroschisis.
one baby treated with primary PC. The third child was born with closed gastroschisis and short-gut syndrome. Given the poor prognosis, the parents initially elected to proceed with palliative care; 1 week later, however, the baby remained stable, and the family elected to proceed with operative intervention, which included a Ladd’s procedure and enteroenterostomy. Silo subset analysis. Thirty-one (39%) babies were treated initially with a silo. Patients treated with a silo required a greater duration of mechanical ventilation (10.64 vs 3.5, P < .001) and accrued greater hospital charges ($249,000 vs $139,000, P < .01). In the PC cohort, 37 (71%) babies were closed primarily, and 15 (29%) required placement of a silo and delayed closure. Sixteen (57%) babies in the SC group were treated with silo and delayed closure. In the silo cohort, PC and SC groups were statistically equivalent when comparing Apgar scores (6.3 ± 2.4 vs 6.2 ± 2.2, P = .78), birth weight (2330 ± 552 vs 2310 ± 608, P = .92), gestational age (254 ± 11 vs 253 ± 13, P = .77), or associated anomalies (5/15 vs 5/16, P = .9). The duration of hospital stay for babies treated initially with silo was 38 ± 14 (days) in the PC group and 50 ± 24 (days) in the SC group (P = .11). Similarly, hospital charges ($266,000 ± $398,000 in the PC group and $232,000 ± $157,000 in the SC group; P = .76), were not statistically different. Babies treated with silo and PC required 8.4 ± 5.4 days of mechanical ventilation, whereas babies treated with silo and SC required 12.9 ± 11.1 days of mechanical ventilation (P = .1; Table IV). Patients required parenteral nutrition for 31 ± 15 days in the PC group and 32 ± 17 days in the SC group (P = .81). Patients started enteral feeds an average of 17 ± 8 days in the PC group and 20 ± 7 days in the SC group (P < .01), which was statistically significant in a multivariate analysis with a 7.5 day advantage with traditional approach.
182 Orion et al
Surgery August 2011
Table II. Multivariate and bivariate analysis of outcomes in babies with gastroschisis Mechanical ventilator (days) SC vs PC (b) Silo vs primary (b) Female GSA (weeks) GI anomaly Cardiac anomaly Second closure Infection Second abdominal surgery Oral aversion Ventilator (days) Reintubation SGS Apgar at 1 min Duration of stay
4.08y 2.08 5.22z 6.64z 2.61* .08 8.32y .94 17.53z 2.40 4.21
9.31* .67
Start enteral feeds (days)x
Parenteral nutrition (days)
Goal enteral feeds (days)x
Infection (OR)
2.67 1.88 1.03 5.36* 3.04
3.08 2.04 4.17 6.13 5.17
2.26 1.56 .84 7.90 7.92
5.15z 2.17 .19* .23y 1.16
1.95 8.72 2.26 3.66
1.57 9.89 33.29* 9.76* 5.30
3.78 10.90 23.78 8.40 5.94
.29
.45* 10.91* 6.32
.53 20.34y 6.07
.37 20.57* 4.64
1.07
Duration of stay (days)
Hospital charges 31,000 ($)x
2.19 7.62 5.7 3.09 0.58 2.95y 9.88 11.77* 9.06 6.32
($35.84) ($6.89) $104.72y $92.63* ($12.25) ($15.67)z $244.88z 48.69 $400.26z ($14.05) ($45.87)
16.34* 1.24y 14.03* 21.79
$109.62 $31.13
.96z
*P < .05. yP < .01. zP < .001. xAnalysis with log transformation for outcomes; effects were calculated as expected change per unit change of independent variables of an average patient in the sample. (b) Bivariate model where only SC vs PC and silo vs primary are controlled. GSA, Gestational age; SGS, short gut syndrome.
Table III. Complications of plastic and SC in gastroschisis patients Complications Any complication Infection/sepsis Hernia Hernia repair Second closure
All 61 32 46 16 3
(76%) (40%) (58%) (20%) (4%)
Plastic 41 22 32 11 1
(79%) (42%) (62%) (21%) (2%)
Although children treated with silo and PC required fewer days to start enteral feeds than those treated with silo and SC, the time to reach goal feeds and discontinue parenteral nutrition was not statistically different. It took an average of 34 ± 14 days in the PC group and 37 ± 15 days in the SC group to reach goal feeds. The average duration in PC and SC groups were 31 ± 15 and 32 ± 17 days, respectively. DISCUSSION The management of gastroschisis remains controversial in part because of long hospital stays and frequent complications associated with this
Suture 20 10 14 5 2
(71%) (36%) (50%) (18%) (7%)
OR [95% confidence interval] .86 1.10 .85 1.15 .51
(.57–1.31) (.80–1.52) (.60–1.19) (.52–2.55) (.21–1.20)
P value .46 .57 .32 .73 .28
anomaly. In the current study, we analyzed the outcomes of babies treated for gastroschisis at a single University Children’s Hospital. The primary goal of the study was to determine the safety and efficacy of PC for abdominal wall defects in gastroschisis. A multivariate analysis was used to compare the PC and SC patient groups. The management approach was based on surgeons’ preference. The patients were candidates for primary closure based on the appearance of the intestines, size of the abdominal wall defect, or hemodynamic status. The patients deemed to be poor candidates for primary closure were treated initially with a preformed, spring-loaded silo and
Orion et al 183
Surgery Volume 150, Number 2
Table IV. Multivariate analysis of outcomes in the silo subset of gastroschisis patients Ventilator (days) Suture vs plastic Female sex GSA Apgar at 1 min GI anomaly Cardiac anomaly Infection Ventilator days Reintubation SGS Second abdominal surgery Second closure Other nonabdominal surgery
7.60* .04 .16 1.06 6.83 3.20 3.16
6.56 18.17* 3.65
Start enteral feeds (days) 7.48z .01 1.21 5.72 .51z 14.11z 21.56z 5.34 9.08
Goal enteral feeds (days)
Parenteral nutrition (days)
2.64 4.17
2.40 7.33
12.30* 10.09 3.95 .67y 17.51z
15.15* 11.77 9.08 .80z 18.49y
16.78z 4.81
16.50y 5.63
*P < .10. yP < .05. zP < .01. SGS, Short gut syndrome.
then underwent either SC or PC once gradual reduction was accomplished. Thirty-one (39%) patients required silo placement as their initial intervention. The findings of this study suggest that both PC and SC might be used to achieve primary abdominal wall closure safely in babies born with gastroschisis. Comparing patients treated with PC and SC using a multivariate statistical analysis, no difference was found between the groups in term of Apgars, gestational age, birth weight, associated anomalies, or comorbidities. The analysis suggests that there was no detectable selection bias between physicians’ preferences of the type of closure, and the 2 treatment groups were comparable. When the outcomes in the entire cohort were analyzed, PC and SC seemed equivalent with regard to duration of hospital stay, time to start enteral feeds, achieve goal enteral feeds, and discontinue parenteral nutrition, as well as hospital charges. The 2 groups differed, however, in the number of days spent on the ventilator and the likelihood of associated infections. The babies treated with PC required lesser duration of mechanical ventilation than babies treated with SC. We speculate that this difference might be related to the lesser intra-abdominal pressures and less impingement on diaphragmatic excursion resulting in better mechanics of breathing in the PC group. The patients treated with PC also had a lower likelihood of having a documented or suspected infection. We speculate that PC allows abdominal drainage, whereas SC is associated
commonly with abdominal wall erythema and systemic inflammatory response. Riboh et al14 reviewed 26 patients treated with PC and analyzed time spent on ventilator, time to initiating enteral feeds, time to discharge from the neonatal intensive care units, and rate of complications. Similar to our findings, the authors demonstrated that sutureless closure was associated with lesser time to extubation. Additional variables that might have contributed to prolonged ventilation in the current study were the presence of GI anomalies, the need for a second abdominal wall closure, low Apgar score at 1 min, and the use of a silo. Interestingly, the presence of cardiac anomalies did not affect the requirement for mechanical ventilation. Prolonged intubation and mechanical ventilation has been shown to increase the risk of tracheal stenosis, ventilator-associated pneumonia, barotraumas, and chronic lung disease in neonates.15,16 Our patients did not sustain complications of prolonged ventilation, but with a mean follow-up of 12 months, we might underestimate the incidence of late adverse events. Weinsheimer et al17 evaluated 99 babies with gastroschisis in the Canadian Pediatric Surgery Network database and identified 9 patients treated using sutureless repair. The authors observed that sutureless closure led to a delay in initiating enteral feeds but interestingly, no increase in the duration of time parenteral nutrition was required and no increase in the duration of stay. We found that the time to initiating enteral feeds was equivalent in our analysis. The difference between the studies might stem from the difference in the
184 Orion et al
approach to initiating enteral feeds, the statistical models used for analysis, and the variables tested. All babies treated with PC, by definition, had a ventral hernia initially, but only 60% had ventral hernias noted either at discharge from the hospital or at their follow-up clinic visit. Surprisingly, only 1 in 5 babies required repair, which was the same as the rate of ventral hernia repair in the SC group. Similar findings were noted for babies treated with silo followed by PC. It seems that the gastroschisis fascial defect contracts and often closes spontaneously in a manner similar to simple pediatric umbilical hernias. Even when ventral herniorrhaphy is required, it could be performed in an elective outpatient setting with minimal morbidity when contrasted with suture repair of gastroschisis in the newborn period. This course in turn might also decrease the cost of the initial hospitalization. The use of preformed silos and the staged reduction of gastroschisis followed by SC have been reported commonly.10,18 Previous studies comparing primary SC with silo with delayed SC suggested that babies in the silo group required a greater duration of mechanical ventilation, hospital stay, and requirement for parenteral nutrition as well as greater hospital charges.12,19,20 Preformed silos can distend the abdominal wall defect, at times requiring a tight closure or the use of prosthetic material to achieve coverage. PC after the placement of a preformed silo has not been reported to date.14 In this study, 15 babies underwent PC after silo placement successfully. The fascial defect in babies treated with PC closed rapidly without delaying discharge from hospital. We compared the outcome of babies treated with silo and PC with babies treated with silo and SC. The traditional staged reduction of gastroschisis requires 2 procedures with the definitive SC performed in the operating room. Transporting ill neonates is time and resource intensive as well as potentially risky. Mobilizing the operating room adds additional drag on constrained hospital and personnel resources. In contrast, silo placement followed by PC, could be performed in the neonatal intensive care unit without a single trip to the operating room.19 The statistical power in the analysis of the silo subset was not sufficient to define the differences in outcome between PC and SC groups because of the relatively small sample size. Babies undergoing PC after silo placement, however, were noted to spend 7.6 fewer days on the ventilator than those undergoing SC. This difference is consistent with our findings for the entire sample of gastroschisis patients. We used a P = .10 as the cutoff for statistical significance because of concern over
Surgery August 2011
sample size, which makes our statement less conclusive. Children treated with silo and PC also required fewer days to start enteral feeds compared with silo followed by SC. Earlier initiation of enteral feeds could decrease the risk of cholestatic liver disease, bacterial overgrowth, and sepsis. No statistical differences were found in the duration of stay, requirement for parenteral nutrition, or time to achieve full enteral feeds, suggesting that gastroschisis associated gastrointestinal dysmotility has greater impact on these outcomes. The treatment of babies with gastroschisis has evolved to a point where long-term survival is expected for most babies. However, a high morbidity is still associated with this anomaly. In our series, 76% of children suffered at least 1 complication. The hospital course is long and expensive with a prolonged period of mechanical ventilation and parenteral nutrition. Our analysis suggests that PC can be a definitive and safe method for treating gastroschisis in babies. PC could contribute to the safety of primary closure and lead to increased rate of primary repair. The procedure can usually be performed in the neonatal intensive care unit in minutes without the need to mobilize the transport team, anesthesia, or the operating room staff. Fewer staged repairs and emergency procedures in the operating room might reduce resource use. In developing countries, where neonatologists, neonatal intensive care units, pediatric surgeons, pediatric anesthesiologists, and mechanical ventilators may not be readily available, PC or silo placement followed by PC might represent a potential approach to decrease the cost as well as the morbidity and mortality of gastroschisis.
REFERENCES 1. Lund SC, Bauer K, Berrios M. Gastroschisis. Incidence, complications, and clinical management in the neonatal intensive care unit. J Perinat Neonat Nurs 2007;2:63-8. 2. Loane M, Dolk H, Bradbury I. EUROCAT working group. Increasing prevalence of gastroschisis in Europe 19802002: a phenomenon restricted to young mothers? Paediatr Perinat Epidemiol 2007;21:363-9. 3. Sekabira J, Hadley GP. Gastroschisis: a third world perspective. Pediatr Surg Int 2009;25:327-9. 4. Sydorak RM, Nijagal A, Sbragia L, Hirose S, Tsao K, Phibbs RH, et al. Gastroschisis: small hole, big cost. J Pediatr Surg 2002;37:1669-72. 5. 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;27:97-101. 6. Howell KK. Understanding gastroschisis: an abdominal wall defect. Neonatal Netw 1998;17:17-25. 7. Canty TG, Collins DL. Primary fascial closure in infants with gastroschisis and omphalocele: a superior approach. J Pediatr Surg 1983;18:707-12.
Surgery Volume 150, Number 2
8. Lacey SR, Carris LA, Beyer AJ III, Azizkhan RG. Bladder pressure monitoring significantly enhances care of infants with abdominal wall defects: a prospective clinical study. J Pediatr Sug 1993;28:1370-4. 9. Dimitriou G, Greenough A, Giffin F, Davenport M, Nicolaides KH. Temporary impairment of lung function in infants with anterior abdominal wall defects who have undergone surgery. J Pediatr Surg 1996;31:670-2. 10. Fischer JD, Chun K, Moores DC, Andrews HG. Gastroshisis: a simple technique for staged silo closure. J Pediatr Surg 1995;30:1169-71. 11. Pastor AC, Phillips JD, Fenton SJ, Meyers RL, Lamm AW, Raval MV, et al. Routine use of a SILASTIC spring-loaded silo for infants with gastroschisis: a multicenter randomized controlled trial. J Pediatr Surg 2008;43:1807-12. 12. Lobo JD, Kim AC, Davis RP, Segura BJ, Alpert H, Teitelbaum DH, et al. No free ride? The hidden costs of delayed operative management using spring-loaded silo for gastroschisis. J Pediatr Surg 2010;45:1426-32. 13. Sandler A, Lawrence J, Meehan J, Phearman L, Soper R. A ‘‘plastic’’ sutureless abdominal wall closure in gastroschisis. J Pediatr Surg 2004;39:738-41.
Orion et al 185
14. Riboh J, Abrajano CT, Garber K, Hartman G, Butler MA, Albanese CT, et al. Outcomes of sutureless gastroschisis closure. J Pediatr Surg 2009;44:1947-51. 15. Srinivasan R, Asselin J, Gildengorin G, Wiener-Kronish J, Flori HR. A prospective study of ventilator-associated pneumonia in children. Pediatrics 2009;123:1108-15. 16. Ho AS, Koltai PJ. Pediatric tracheal stenosis. Otolaryngol Clin North Am 2008;41:999-1021. 17. Weinsheimer RL, Yanchar NL, Bouchard SB, Kim PK, Laberge J, Skarsgard ED, et al. Gastroschisis closure- does method really matter? J Pediatr Surg 2008;43:874-8. 18. Lansdale N, Hill R, Gull-Zamir S, Drewett M, Parkinson E, Davenport M, et al. Staged reduction of gastroschisis using preformed silos: practicalities and problems. J Pediatr Surg 2009;44:2126-9. 19. McNamara WF, Hartin CW, Escobar MA, Lee Y. Outcome differences between gastroschisis repair methods. J Surg Res 2011;165:19-24. 20. Bayard D, Ramones T, Phillips SE, Leys CM, Rauth T, Yang EY. Method to our madness: an 18-year retrospective analysis on gastroschisis closure. J Pediatr Surg 2010;45: 579-84.