Gastroschisis: A Sixteen-Year Review

By J. Baerg, G. Kaban, J. Tonita, P. Pahwa, and D. Reid Saskatchewan, Canada and Loma Linda, California

Purpose: The aim of this study was to examine the incidence of gastroschisis in Saskatchewan over a 16-year period and identify risk factors associated with mortality. Methods: Information was obtained by retrospective review of all infant charts at the only 2 provincial neonatal intensive care centers from January 1985 through December 2000. Factors recorded were gestational age, birth weight, gender, Apgar score, time to operation, method of closure, time to attain full feedings, presence of sepsis, and length of hospital stay. Live birth data for the province was obtained, and the incidence per 10,000 live births for 3 time intervals was calculated. Analysis was performed with mortality as the dependent variable. Results: Seventy-one infants were identied. Overall survival rate was 93% (66 of 71). During the 3 time intervals examined, the incidence of gastroschisis per 10,000 live births

increased from 1.85 in 1985 to 1990 to 3.66 in 1991 to 1995 to 4.06 in 1996 to 2000. The analysis found that intestinal atresia P .009, OR 18.3 (95%CI: 2.457136) and intestinal necrosis P .050, OR 10.33 (95%CI: 1.32 80.68) were signicantly associated with mortality. Conclusions: The incidence of gastroschisis is increasing. Intestinal atresia and the development of intestinal necrosis were associated with poor outcome. Patients who had these complications all had short bowel, and full feedings could not be established. They required continuation of total parenteral nutrition and experienced episodes of sepsis. J Pediatr Surg 38:771-774. 2003 Elsevier Inc. All rights reserved. INDEX WORDS: Gastroschisis, intestinal atresia, short bowel syndrome.

ASTROSCHISIS is increasing throughout the world for unknown reasons.1 This condition now occurs in about 1 of every 4,000 live births.2 Saskatchewan is a province in western Canada with a consistent population of just over one million people and a declining birth rate.3 We examined the incidence of gastroschisis and compared it with the live birth rate in the province to nd out if the condition also is increasing in this relatively xed population. As neonatal care and parenteral nutrition techniques have improved, the survival rate of gastroschisis patients has correspondingly improved. Modern centers now expect survival rates of over 90%.4 We examined our data to compare our survival rate with that of other centers and identify any factors associated with poor outcome.
MATERIALS AND METHODS
Information was obtained by retrospective review of all infant and maternal charts at the only 2 provincial neonatal intensive care centers from January 1985 through December 2000. All charts coded for the condition of gastroschisis were retrieved. Factors recorded for the analysis included birth weight, gestational age, gender, Apgar score, mode of delivery, mode of defect closure, time to arrive in the operating room, presence of sepsis dened as positive blood cultures, time to attain full feedings, and length of hospital stay. The presence of intestinal atresia, perforation, or intestinal necrosis was recorded. Mortality rate was dened as death during the initial hospital stay. Live birth data for the province were obtained. The incidence of gastroschisis per 10,000 live births for each interval was calculated. The incidence was compared with the provincial live birth rate for each corresponding interval.3 Relative risk and 95% condence intervals for each period were calculated. Characteristics of survivors and nonsurJournal of Pediatric Surgery, Vol 38, No 5 (May), 2003: pp 771-774

vivors were compared. The Fishers Exact test was used for proportional analysis to calculate the P value. P values of less than .050 were considered signicant. Odds ratios and 95% condence intervals were calculated as an estimate of relative risk with death as the dependent variable.

RESULTS

During the 16 years examined, all patients were born within the province. None were transferred into the province for care after birth. The average maternal age was 21 years (range, 14 to 45 years), and 61% (43 of 71) were primigravida. To nd out if the incidence of gastroschisis was actually increasing, the 16-year data collection period was divided into 3 time intervals (Table 1). The incidence increased from 1.85 per 10,000 live births between 1985 and 1990 to 4.06 per 10,000 live births between 1996 and 2000. During this 16-year period, the number of live births per 1,000 population fell from 18.1 to 12.6.3 The relative risk of a child being born with gastroschisis
From Regina General Hospital, University of Saskatchewan, Saskatchewan, Canada, and Childrens Hospital, Loma Linda, CA. Presented at the 34th Annual Meeting of the Canadian Association of Paediatric Surgeons, Vancouver, British Columbia, Canada, September 19-22, 2002. Address reprint requests to Joanne Baerg, MD, Loma Linda University, Room 21111, Coleman Pavilion, 11175 Campus St, Loma Linda, CA 92354. 2003 Elsevier Inc. All rights reserved. 0022-3468/03/3805-0028$30.00/0 10.1016/S0022-3468(03)00008-3
771

772

BAERG ET AL

Table 1. Incidence and Relative Risk

Time Period Live Births per 1,000 Population Incidence per 10,000 Live Births Relative Risk and 95% CI

1985-1990 1991-1995 1996-2000

18.1 14.5 12.6

1.85 3.66 4.06

1.0 1.94 (1.07-3.50) 2.24 (1.25-4.03)

in the 1996 to 2000 time interval was 2.24 (95%CI 1.25 4.03). This doubling was statistically signicant. Table 2 summarizes the characteristics of the patients in the database. There were 71 patients, 36 boys (51%) and 35 girls (49%). The average birth weight was 2,519 g, and the average gestational age was 36.7 weeks. Average Apgar scores were 6.2 at 1 minute and 8 at 5 minutes. Forty patients (56%) were born by vaginal delivery, and 31 (44%) were delivered by cesarian section. Six patients underwent cesarian section because of the diagnosis of gastroschisis. Seventeen cesarian sections were performed for fetal indications, and 3 were for maternal indications. In 5 cases, the indication for cesarian section is unknown. Primary closure was carried out for 56 patients (79%) and 15 (21%) were managed with silo closure. The average time from birth to surgery was 6.8 hours (range, 1.3 to 15.7 hours). The average time for all patients to tolerate full feedings and no longer require total parenteral nutrition (TPN) was 35 days (range, 7 to 356 days). The average hospital stay was 50 days (range, 8 to 356 days). The overall survival rate was 93% (66 of 71). Patients managed with silo closure were examined. Eighty-seven percent (13 of 15) survived. The average age of those managed with a silo was greater than those managed with primary closure, 36.9 weeks versus 36.8 weeks, respectively. The average birth weight, however, was lower, 2,461g versus 2,516 g. No complications occurred directly from the silo, although one patient had intestinal necrosis, and 2 presented with intestinal atresia. Seven patients (45%) had episodes of sepsis. The average hospital stay was almost twice as long for those managed with silo closure, 77.8 days compared with 41.9 days for those managed with primary closure (independent t test; P .160). The average amount of time to achieve full feeding after silo closure was 54.7 days compared with 28.5 days for patients who underwent primary closure (independent t test; P .295). The difference between the variables was not statistically signicant. Characteristics of survivors and mortalities were compared (Table 3). Nonsurvivors were not of signicantly lower birth weight or gestational age. Only 13% (9 of 71) of patients were born before 35 weeks gestational age, and all of these patients survived. All 5 patients who died weighed over 2,400 g and were born after 35 weeks

gestational age. Gender, mode of delivery, and mode of closure were not factors associated with mortality. The average time to arrive in the operating room was 6.7 hours for survivors and 8.3 hours for those who died (independent t test; P .741). Babies who died had higher 1-minute and 5-minute average Apgar scores, 7.6 versus 6.2 and 9 versus 8, respectively. None of these factors involving early management were signicantly associated with mortality. Intestinal atresia was present in 8 cases (11%). The presence of intestinal atresia was associated with mortality (P .009, OR 18.3, 95% CI: 2.46 136). Two cases involved colon, 2 involved small bowel, and 4 had multiple atresias involving both large and small bowel. All patients with intestinal atresia underwent surgery to place the bowel in continuity. Five patients who survived had adequate gut length and did not require long-term TPN. All were discharged from the hospital tolerating full feedings. Their average time to achieve full feedings was 41.8 days, and average hospital stay was 86 days. Three patients had short gut with TPN dependence. All 3 died of liver failure related to short gut, TPN, and recurring sepsis episodes. Full feedings were never established, and TPN continued to be required. The average time to death for these 3 was 200 days. Intestinal perforation was present in 3 patients (4%) at the time of delivery. All 3 cases were managed initially with resection, and the bowel was placed in continuity. Two underwent primary closure, and one was managed with a silo, and no complications occurred related to mode of closure. The presence of bowel perforation at the time of gastroschisis presentation was not a statistically signicant factor associated with mortality (P .199, OR 8.00, 95% CI: 0.59 108.17). Intestinal necrosis developed during the initial hospital stay in 6 patients (8%). All underwent resection of the
Table 2. Characteristics of Patients With Gastroschisis (n 71)
Range

Mean birth weight (g) Mean gestational age (wk) Gender M/F Mean Apgar at 1 min Mean Apgar at 5 min Vaginal delivery Cesarian delivery Primary closure Silo closure Intestinal atresia Perforation at birth Necrosis Mean time on TPN (d) Mean hospital stay (d) Mean time to OR (h) Sepsis Mortalities

2,519 36.7 36 (51%)/35 (49%) 6.2 8 39 (55%) 31 (45%) 56 (79%) 15 (21%) 8 (11%) 3 (4%) 6 (8%) 35 50 6.8 27 (38%) 5 (7%)

1,100-4,200 32-41 1-10 2-10

7-356 8-356 1.3-15.7

GASTROSCHISIS Table 3. Comparison of Survivors and Mortalities (n 71)

Survivors (66, 93%) Mortalities (5, 7%) P Value Odds Ratio (95%CI)

773

Birth Weight Gestational age Gender M/F Delivery V/C Closure P/S Sepsis Atresia Perforation Necrosis

2,514 g 36.8 wk 33/33 36/30 53/13 23 5 2 4

2,583 g 36.6 wk 3/2 4/1 3/2 4 3 1 2

.954 .820 1.00 .378 .283 .066 .009 .199 .050

NA NA 1.50 (0.24-9.57) 3.33 (0.35-31.44) 0.37 (0.06-2.43) 7.48 (0.79-70.88) 18.3 (2.46-136.00) 8.00 (0.59-108.17) 10.33 (1.32-80.68)

Abbreviations: V/C, vaginal/cesarean; P/S, primary/silo.

nonviable gut. Two of these patients died, and, when the analysis was carried out, necrosis was found to be associated with mortality (P .050, OR 10.33, 95% CI: 1.32 80.68). Twenty-seven patients (38%) had sepsis during the initial hospital stay. Examination of these patients found a trend toward mortality that was not statistically significant (P .066, OR 7.48, 95% CI: 0.789 70.884). All babies who died had evidence of intestinal atresia, perforation, or necrosis. Intestinal atresia and necrosis were statistically signicant factors in this study. The average time to establish full feedings in survivors was 29 days, and the average hospital stay was 45 days. For those who died, the average hospital stay was 125 days, and full feedings were never established. Thirty-one infants (44%) were breastfed successfully, and all but one survived. One child, who initially attained full feedings and was discharged, died at 7 months of age from SIDS.
DISCUSSION

Despite a drop in live birth rate in the province, the incidence of gastroschisis doubled over the 3 time intervals examined in this study (Relative risk 2.24, 95%CI: 1.25 4.03). This nding corresponds to a worldwide increase in the incidence of the condition (Table 4). The cause of the increase has not been identied despite multiple retrospective studies that examine potential teratogens and maternal demographics.8 The average maternal age for the patients examined in this study was 21. This is similar to that of other published series.9 The overall survival rate of patients in this series was 93% (66 of 71). Other investigators report similar surTable 4. The Increasing Incidence of Gastroschisis
Country Period of Study Increase per 10,000 Live Births

North England5 Southwest England6 Japan1 Hawaii7

1986-1996 1987-1995 1975-1997 1986-1997

1.48-4.72 1.60-4.40 1.31-4.67 2.52-3.85

vival rates.4 With current neonatal care and nutritional techniques, one can expect the mortality rate for gastroschisis to remain under 10%. All 5 patients who died in this series had intestinal atresia, perforation, or intestinal necrosis. When the database was analyzed for factors associated with poor outcome, only intestinal atresia and necrosis achieved signicance. All 5 had ongoing problems related to the fact that full feedings could never be established. They required TPN for the entire hospital stay and complications related to short bowel syndrome, liver failure, and multiple septic episodes led to their eventual death. Sepsis was of borderline signicance for mortality (P .066), and this may reect a small sample size. In his series of 185 patients, Snyder10 reported that 83% (15 of 18) of patients with intestinal atresia and gastroschisis survived. Although patients with atresia had delayed feeding and prolonged hospital stays, there was no correlation between atresia and mortality rate. Driver11 reports that intestinal atresia was not an independent risk factor for mortality in his series of 91 patients. Other investigators have identied intestinal atresia as a marker of increased morbidity and mortality.12 Snyder10 found that intestinal necrosis was an independent predictor of increased mortality. Any sign of intestinal ischemia in an infant with gastroschisis should be managed with bowel rest, antibiotics, serial radiographs, and clinical and laboratory examinations. Birth weight, gestational age, and mode of delivery did not correlate with mortality in this study. Only 9 patients (13%) were born before 35 weeks gestational age, and all survived. All 5 patients who died weighed over 2,400 g at birth and were over 35 weeks gestational age. These ndings reect reports of other investigators who emphasize early delivery by elective cesarian section may produce premature infants.10 The risks of prematurity may outweigh the risks associated with gastroschisis, which should have a survival rate of more than 90%.4

774

BAERG ET AL

Mode of closure did not correlate with mortality. Seventy-nine percent (56 of 71) of infants in this series underwent uneventful primary closure. Snyder10 reports that 75% of 185 patients in his series were treated successfully with primary closure. Patients who underwent primary closure attained full feedings in less time than those closed with a silo. The earlier manifestation of gut function may reect the fact that these patients were able to undergo abdominal closure without respiratory, gastrointestinal, or renal compromise. Some investigators now report mitigation of intestinal necrosis in patients who undergo silo closure.13 Because intestinal necrosis is a factor

associated with mortality, we would advocate silo use whenever any risk of intestinal compromise is suspected. Gastroschisis is increasing, and our data reect this worldwide unexplained increase (Table 4). Intestinal atresia and intestinal necrosis were associated with poor outcome. All infants who died had short gut and a corresponding inability to establish full feedings. The continuation of TPN and subsequent liver failure, as well as multiple episodes of sepsis related to central venous lines, contributed to morbidity and mortality in these patients. An overall survival rate of 93% is similar to those in other published series.

REFERENCES
1. Suita S: Changing prole of abdominal wall defects in Japan: Results of a national survey. J Pediatr Surg 35:66-72, 2000 2. Tunell WP: Omphalocele and gastroschisis, in Ashcraft KW, Holder TM, (eds): Pediatric Surgery. ed 2. Philadelphia, PA, Saunders, 1993, pp 546-556 3. Statistics Canada, Population Statistics for Saskatchewan 19952000 4. Novotny DA, Klein RL: Gastroschisis: An 18 Year Review. J Pediatr Surg 28:650-652, 1993 5. Rankin J: Congenital anterior abdominal wall defects in the north of England, 1986-1996: Occurrence and outcome. Prenat Diagn 19: 662-668, 1999 6. Penman DG: Increase in incidence in gastroschisis in the southwest of England in 1995. Br J Obstet Gynecol 106:328-331, 1998 7. Forrester M: Epidemiology of abdominal wall defects, Hawaii, 1986-1997. Teratology 60:117-123, 1999 8. Torfs C: Maternal medications and environmental exposures as risk factors for gastroschisis. Teratology 54:84-92, 1996 9. Goldbaum G: Risk factors for gastroschisis. Teratology 42:397403, 1990 10. Snyder C: Outcome analysis for gastroschisis. J Pediatr Surg 34:1253-1256, 1999 11. Driver CP: The contemporary outcome for gastroschisis. J Pediatr Surg 35:1719-1723, 2000 12. Cusick E: Small-bowel continuity: A crucial factor in determining survival in gastroschisis. Pediatr Surg Int 12:34-37, 1997 13. Fischer J: Gastroschisis: A simple technique for staged silo closure. J Pediatr Surg 30:1169-1171, 1995