The Use of a Spring-Loaded Silo for Gastroschisis: Impact on Practice Patterns and Outcomes

Objective  To evaluate the impact of the use of a bedside-placed spring-loaded silo (SLS) on practice patterns and on outcomes for infants with gastroschisis.

Design  Retrospective review comparing neonates with gastroschisis treated before and after the implementation of selective SLS placement.

Setting  Tertiary referral center.

Patients  Of 91 consecutive neonates admitted for initial treatment of gastroschisis between January 1998 and August 2007, 45 were admitted before and 46 were admitted after implementation of the SLS.

Main Outcome Measures  Immediate fascial closure rate, infection rate, time to fascial closure, time to initiation of enteral feeding, time to achievement of full enteral feeds, time of hyperalimentation requirement, and length of hospital stay.

Results  The rate of immediate fascial closure was lower in the postimplementation group (58% before vs 20% after implementation, P < .001). Overall length of stay, time to enteral feeding, and infection rates were not significantly different between the 2 groups.

Conclusions  The use of an SLS placed at the bedside has resulted in lower immediate fascial closure rates for infants with gastroschisis without significant detrimental clinical outcome. The main benefit of using the bedside-placed SLS is the avoidance of urgent surgical intervention. For patients undergoing delayed fascial closure, use of the bedside SLS resulted in shorter times to definitive fascial closure.

Gastroschisis is a congenital anomaly of the abdominal wall that has historically been treated on an emergency basis by primary closure or, in the case of abdominovisceral disproportion, by surgical silo placement. Ongoing controversy exists regarding the optimal surgical treatment of this anomaly. Bedside placement of a spring-loaded silo (SLS) (Ventral Wall Defect Silo Bags; Bentec Medical, Woodland, California; Figure 1) was first described in 1995 and was implemented at our institution in January 2004.¹ Proposed benefits of this device have included fewer days in need of ventilatory support, decreased incidence of pulmonary barotrauma, shorter time to enteral feeding, improved tissue perfusion, improved cosmetic outcome, decreased incidence of infectious complications, avoidance of emergency surgical intervention, and lowered hospital charges owing to shorter stay and fewer complications.^1-5 Several studies have described an initial experience with this device,^1,6 including a retrospective comparison of selected cases of intermittent SLS placement^3,4 or routine SLS placement^2,5 vs urgent surgical treatment.

The choice of immediate surgical therapy or bedside-placed SLS in our institution has been determined in most cases by surgeon discretion. Therefore, direct comparison of SLS with traditional treatment is subject to selection bias, and the impact of the SLS is most appropriately evaluated within the context of overall patient outcomes with selective use. To our knowledge, there are no data regarding the impact of the availability of this device on practice patterns or on patient outcomes because previous study designs do not account for potential selection bias in treatment modality. The objective of this study was to investigate changes in practice patterns and potential changes in patient outcomes related to the availability and selective use of the SLS in our institution. Therefore, comparisons are made between overall patient outcomes, regardless of treatment modality, before and after selective implementation of the SLS.

Medical records were reviewed for all patients admitted to Seattle Children's Hospital for initial treatment of gastroschisis between January 1998 and August 2007. Patients were identified on the basis of an operating room scheduling database and by International Classification of Diseases, Ninth Revision, diagnostic codes.⁷ Patients were excluded if they underwent surgical repair at an outside facility before presentation to our institution.

Initial surgical treatment was considered in 3 distinct categories: (1) immediate fascial closure, (2) sutured silo with delayed fascial closure, and (3) SLS with delayed fascial closure. Immediate fascial closure referred to definitive fascial closure performed within the first 6 hours of life without prior treatment with SLS or sutured silo. Sutured silo with delayed fascial closure referred to primary placement of a formal silicone rubber (Silastic; Dow Corning, Midland, Michigan) silo by suturing it to the fascia through an extended midline incision. And SLS with delayed fascial closure referred to the placement of an SLS at the bedside or in the operating room within the first 6 hours of life. Delayed fascial closure was then attempted 2 to 5 days later. All patients were classified according to the first method of treatment (ie, SLS with subsequent conversion to sutured silo due to increased intra-abdominal pressures (IAPs) at the time of attempted fascial closure was classified as SLS with delayed fascial closure). Definitive fascial closure was defined as fascial closure by suture or by tension-free methods, including the use of prosthetic mesh or patch.

Before implementation of the SLS, all patients were treated with urgent surgical intervention. All neonates arrived at our institution via transfer from a referring hospital, because our freestanding children's hospital does not have a labor and delivery ward. Because of the delay in treatment due to transfer, surgical treatment was attempted as soon as possible after arrival. Immediate fascial closure was attempted in all patients with real-time IAP measurement as determined by intragastric pressure monitoring. Neonates with IAPs greater than 20 mm Hg were treated with a midline incision and a reinforced silicone rubber silo sutured to the fascia. Patients were then transferred to the intensive care unit and continued to receive mechanical ventilation. Silo reduction with the use of a wringer clamp was performed once or twice daily as previously described.⁸ Once the visceral contents were reduced to the skin level, patients were taken to the operating room for definitive fascial closure.

Initially after the implementation of the bedside-placed SLS, patients either underwent urgent surgical treatment or received a bedside-placed silo, per surgeon preference. Some patients with confounding factors such as skin bridges or omphalomesenteric bands were taken to the operating room for SLS placement.⁹ Patients treated with SLS did not receive mechanical ventilation unless indicated for unrelated causes (eg, acute respiratory distress syndrome, aspiration). Visceral contents were allowed to reduce under gravity, and delayed fascial closure was performed in the operating room under general anesthesia with IAP monitoring. Spring-loaded silo treatment was generally allowed to proceed for up to 5 days before surgical intervention (attempted fascial closure or formal sutured silo placement depending on the IAP). Because treatment with the SLS has been subjectively successful over time, practice patterns have evolved to routine use of the device.

This was a retrospective study. Approval from the Seattle Children's Hospital institutional review board was obtained before review of the patient records. Patients who were seen through December 31, 2003 (45 neonates) were considered a preimplementation (historical) control group, whereas patients seen between January 1, 2004, and August 2007 (46 neonates) were considered the postimplementation group. Practice pattern outcomes included immediate fascial closure rates. Patient outcomes included infection rate, time to fascial closure, time to initiation of enteral feeding, time to achievement of full enteral feeds, time of hyperalimentation requirement, and length of hospital stay. Patients were excluded from length of stay and time to enteral feeding analysis if they had a confounding condition including atresia, short gut syndrome at the time of discharge, stricture, web, perforation, severe gastroesophageal reflux, or an associated congenital anomaly that prolonged care (1 infant with total anomalous pulmonary venous return) (Table 1) or if they underwent regional transfer to an outside facility before cessation of hyperalimentation (3 infants).

Outcome data were significantly nonnormal and therefore were analyzed according to nonparametric methods (Mann-Whitney test for continuous data and χ² test of independence for frequency data). To control for experiment-wide type I error (α = .05), a Bonferroni correction for 17 comparisons was used with P < .003 per comparison considered statistically significant. For χ² analysis involving frequency counts of less than 5, a conservative Yates correction was used.

Baseline patient characteristics were similar in the 2 groups (Table 1). Overall mortality was 1%. One neonate in the postimplementation group died of sepsis at 4 days of life and was excluded from all analyses. Primary closure rates were significantly lower after implementation of the SLS (58% vs 20%, P < .001) (Figure 2). Abdominal compartment syndrome was not observed in either group. Intra-abdominal pressure at time of closure was similar in the 2 groups (mean, 10 vs 11 mm Hg for the preimplementation group vs the postimplementation group; P = .37). Time to fascial closure was not significantly different between groups (mean, 5.3 vs 5.7 days; P = .30). Among patients treated with bedside-placed SLS, successful fascial closure was achieved in 90% (19 of 21) of patients on the first attempt. Two patients required conversion to a formal sutured silo owing to persistent abdominovisceral disproportion; 1 of the 2 eventually required closure with bioprosthetic mesh. Among patients undergoing SLS placement in the operating room, closure was successful in 75% (6 of 8) of the subsequent operating room trips. Reasons for failure to achieve fascial closure included elevated IAP and perforated atresia.

Infectious complications were not significantly different between groups, but a possible trend was noted for decreased overall infectious complications in the postimplementation group (Table 2). No significant differences were seen in length of stay, time to initiation of enteral feeding, time to achievement of full enteral feeds, time to advance enteral feeding to goal rate (as calculated by the interval between initiation of enteral feeding and achievement of full feeds), or time of hyperalimentation requirement (Table 3).

This study is the first to our knowledge to demonstrate the impact of the selective implementation of an SLS on practice patterns and subsequent clinical outcomes. A significant decrease was seen in immediate fascial closure rate, but clinical outcomes, including length of stay and time to enteral feeding, were not negatively affected. It is difficult to make a definitive statement about infectious complications other than that the demonstrated postimplementation wound infection rate is consistent with that reported in the literature. The lack of significance in overall infection rate may represent a type II error due to small numbers or may be due to difficulty in defining and determining a true infectious complication. The definition of infectious complication that was used for data abstraction was documented erythema requiring the initiation of antibiotics at any point after fascial closure. In the control group, however, patients routinely had abdominal wall erythema immediately after initial primary closure and were frequently treated with an extended course of antibiotics, making subsequent determination of infection difficult. Therefore, the rate of infection is likely underreported in the preintervention group because most of these children could not be characterized by the definition used.

Previous studies have shown a mixed impact on time to achievement of full enteral feeds and time to hospital discharge, possibly related to patient selection criteria. Our data suggest that these 2 outcomes are not detrimentally affected by the selective implementation of the SLS. In addition, outcomes of infections have demonstrated mixed results, likely due to small sample sizes and the previously mentioned difficulty in defining wound infections.

The retrospective nature of this study limits the definitive conclusions of this study, but at this point, it represents the best available evidence for the implementation of the SLS. Other factors, such as hand washing, antibiotic use, and central venous access procedures, have changed with time and may affect outcomes. The small number of patients in this single-institution study limits the generalizability of these findings, and further multicenter prospective study is warranted to reinforce these findings. One of the major proposed benefits of the bedside-placed SLS is the decreased need for mechanical ventilation. While this study did not analyze ventilatory parameters, our experience has been that patients treated with SLS generally do not require mechanical ventilation unless indicated for causes not related to the abdominal wall defect. Use of the SLS also changes the nature of initial surgical treatment from emergency to semielective. Many neonates develop some degree of intestinal edema in the time it takes to transport them to our institution. The use of the SLS allows this edema to resolve and may increase our rate of primary facial closure (although in a delayed fashion) and in doing so avoid placement of a midline sutured silo, resulting in a long-term cosmetic benefit.

The availability and implementation of the SLS at our institution have led to a decrease in immediate fascial closure rates without a demonstrated detrimental effect on patient outcomes as measured by length of stay, time to enteral feeding, and infectious complications. This study suggests that neonates with gastroschisis can be safely treated initially with an SLS placed at the bedside, thereby not requiring an urgent surgical intervention. The avoidance of urgent surgical intervention may allow for medical optimization, including hemodynamic stabilization, diuresis, and resolution of intestinal edema before definitive closure. A randomized longitudinal study is warranted to validate these hypotheses.

Correspondence: Stephen S. Kim, MD, Division of General and Thoracic Surgery, Seattle Children's Hospital, 4800 Sand Point Way, Mailstop W7729, Seattle, WA 98105 (stephen.kim@seattlechildrens.org).

Accepted for Publication: April 7, 2008.

Author Contributions: Drs Jensen, Waldhausen, and Kim had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Jensen and Kim. Acquisition of data: Jensen. Analysis and interpretation of data: Jensen, Waldhausen, and Kim. Drafting of the manuscript: Jensen. Critical revision of the manuscript for important intellectual content: Waldhausen and Kim. Statistical analysis: Jensen. Study supervision: Waldhausen, Kim.

Financial Disclosure: None reported.

Previous Presentation: This study was presented at the Annual Meeting of the Pacific Coast Surgical Association; February 16, 2008; San Diego, California.