OBJECTIVE: We sought to compare neonatal survival of infants with gastroschisis by mode of delivery.
METHODS: We conducted a retrospective cohort study on infants with gastroschisis who were delivered in New York State from 1983 through 1999. We compared neonatal mortality between infants born vaginally and those delivered by cesarean using adjusted hazard ratios derived from Cox proportional hazards regression models.
RESULTS: A total of 354 infants were found to have isolated gastroschisis. Of these, 174 were delivered vaginally, whereas 180 were delivered by cesarean. Neonatal mortality was registered among 18 infants (5.1%); 12 (6.9%) in the vaginal and 6 (3.3%) in the cesarean group. After controlling for potential confounders, the risk for neonatal demise was similar in both the vaginal and cesarean subcohorts (adjusted hazard ratio 0.84, 95% confidence interval [CI] 0.29–2.43). Preterm birth was the morbidity pathway that explained the early demise of infants with gastroschisis, irrespective of mode of delivery (adjusted hazard ratio 3.4, 95% CI 1.10–10.4) whereas small for gestational age did not predict mortality (adjusted hazard ratio 1.04, 95% CI 0.13–8.14).
CONCLUSION: In this study the mode of delivery was not found to be associated with neonatal survival of infants with gastroschisis. Preterm birth rather than small for gestational age was the predictor of neonatal death among gastroschisis infants.
LEVEL OF EVIDENCE: III
Abdominally and vaginally delivered infants with gastroschisis seem to have equal likelihood of neonatal survival.
From the *Department of Maternal and Child Health, University of Alabama at Birmingham; †Congenital Malformations Registry, New York State Department of Health, Troy, New York; ‡Department of Biostatistics and Epidemiology, University of South Florida, Tampa; and §Department of Internal Medicine, Saint Agnes Hospital, Baltimore, Maryland.
Partly supported through a Young Clinical Scientist Award to the first author (H.M.S.) by the Flight Attendant Medical Research Institute (FAMRI).
Address reprint requests to: Hamisu Salihu, MD, PhD, Department of Maternal and Child Health, University of Alabama at Birmingham, 1665 University Boulevard, Room 320, Birmingham, AL 35294; e-mail: firstname.lastname@example.org.
Received April 18, 2004. Received in revised form June 11, 2004. Accepted June 17, 2004.
Recently, there has been a growing interest in gastroschisis, mainly because of 2 phenomena: first, using different study designs and analytic approaches, researchers1–9 have noted a progressive rise in the incidence of gastroschisis in several populations, and second, researchers1–3 have consistently noted that gastroschisis tends to cluster among younger women. Despite these differences, the obstetric management of gastroschisis and omphalocele are similar,10 although obstetricians are at odds regarding the optimal mode of delivery of infants with abdominal wall defects.
The inconsistency of results across studies as to the safer route of delivery (vaginal versus cesarean) may be the result of the small sample size, the inability to control for confounding characteristics, and selection biases induced mainly by differences in referral rates and patterns.
In this study we analyzed population-based data from the New York State Congenital Malformations Registry, one of the largest birth defects registries in the nation, with the following hypotheses: 1) infants with gastroschisis who are delivered abdominally have better chances of survival than their vaginally delivered counterparts and 2) irrespective of the route of delivery, prematurity and small for gestational age are independent predictors of neonatal survival among isolated cases of gastroschisis.
MATERIALS AND METHODS
We obtained data for this study from the New York State Congenital Malformations Registry for the surveillance period from the calendar year 1983 through 1999. The New York State Congenital Malformations Registry uses passive case ascertainment, relying on reports from hospitals and physicians. State regulations mandate all physicians and other hospital staff to report major congenital malformations at birth through the age of 2 years. New case reports are matched against existing registry data to eliminate duplication. For each case, reporting physicians, hospitals, and genetics laboratories are asked to provide a narrative description of the congenital anomaly, and the staff at New York State Congenital Malformations Registry reviews all such reports carefully. Incomplete reports and nonspecific diagnoses are followed up with the reporting source.
Most New York State Congenital Malformations Registry prevalences for major birth defects are similar to those of other registries that use active case finding, such as the California Birth Defects Monitoring Program and the Metropolitan Atlanta Congenital Defects program.11 The completeness of the New York State Congenital Malformations Registry data has been found to be satisfactory by other investigators.12
On the basis of the diagnosis and narrative information provided by the reporting entity, we defined gastroschisis as an abdominal wall defect located laterally to a normal umbilicus but not involving the umbilical ring. The herniated viscera often are covered by exudate, but no membrane covers the loops of gut floating in amniotic fluid.
The primary outcome of interest was neonatal mortality, defined as the death of the child within the first 28 days of life. Information on neonatal mortality was obtained through the routine linkage conducted by the New York Vital Records Office as required by the National Center for Health Statistics. We compared neonatal mortality between infants with gastroschisis who were born vaginally and those who were delivered by cesarean. We also compared the distribution of fetal growth parameters between the 2 groups; low and very low birth weight (< 2,500 and < 1,500 g), preterm and very preterm delivery (< 37 and < 33 weeks of gestation), postterm (≥ 42 weeks of gestation), and small for gestational age (< 10th percentile of birth weight for gestational age) using US fetal growth reference curves for singletons.13 This analysis was performed to measure the level of fetal growth morbidity among affected infants and to assess to what extent infant survival was compromised in the presence or absence of these negative fetal growth indices (fetal growth and shortened gestation), stratified by mode of delivery.
In a comparative analysis to provide background information on the study cohort, we generated a random sample of approximately 1 million anomaly-free singleton live births that occurred in the State of New York using the Vital Statistics records. We then compared indices of fetal growth parameters (low and very low birth weight, preterm and very preterm, postterm, and small for gestational age) between these anomaly-free infants and those identified with gastroschisis. We also compared the vaginal and abdominal groups with respect to the following sociodemographic characteristics: maternal age, race/ethnicity, parity, maternal education, place of residence, mode of delivery, and level of care. Maternal age was categorized as younger than 25 years and 25 years or older based on our previous report demonstrating the preponderance of gastroschisis among younger mothers.1 Parity was classified into nullipara (< 1) and multipara (≥ 1). Maternal education was subdivided into mothers with a high school level of education and those mothers with less than a high school education. Place of residence was categorized into living either in New York City or Upstate New York. The level of perinatal care was classified into lower, representing level of care where specialized surgery is not offered, and higher, composed of health facilities in which specialized surgery is conducted.
In this study, only isolated cases of gastroschisis were considered. If the abdominal defect was the only anomaly in the infant, then it was isolated. Infants with minor congenital anomalies (eg, isolated polydactyly) or gastrointestinal anomalies associated with the ventral wall defect also were classified as isolated.14 The reason for limiting our analysis to isolated cases is that multiple malformed gastroschisis may in fact represent a different disease variant with significantly higher mortality, and interpretations drawn from analysis on an admixture of the 2 forms of gastroschisis will lead to misleading conclusions.
Before the beginning of analysis, we obtained approval from the New York State Department of Health Institutional Review Board. We conducted our analyses using the SAS software (9.0; SAS Institute, Cary, NC). We conducted crude comparisons by means of 2 by 2 contingency tables and used the χ2 test to estimate differences in proportions. Neonatal mortality rate among infants with gastroschisis was calculated by dividing counts of neonatal deaths by total number of live births within the period multiplied by 100. We used the Cox proportional hazards regression model to derive adjusted hazard ratios for neonatal mortality after testing for the nonviolation of the proportionality assumption in each case.15 We confirmed this by plotting the log-negative-log of the Kaplan–Meier estimates of the survival function versus the log of time. The resulting curves were parallel. Adjusted hazard ratios were derived by loading all the variables considered potential confounders into the model. We also conducted power analysis to determine sample size requirements under the following assumptions: 1) a type 1 error rate of 5%, 2) 80% power, 3) a relative risk in neonatal mortality that varies from about 2 to 4 times as great between abdominally and vaginally delivered babies with gastroschisis, and 4) an equal sample size in each arm. The expected sample size in each arm (vaginally and abdominally delivered infants) for the different relative risks for neonatal mortality is illustrated in Figure 1. All tests of hypothesis were 2-tailed with a type 1 error rate set at 5%.
Three hundred fifty-four isolated cases of gastroschisis were analyzed, of which 174 (49.2%) were delivered vaginally and 180 (50.8%) were delivered by cesarean. Mothers in both groups were similar in terms of age, parity, race/ethnicity, educational level, and place of residence (Table 1). Classification of facility where delivery took place was documented in 349 of the 354 cases (98.6%), and the majority of the deliveries (76.8%) were conducted at specialized centers. However, a significant difference was noted between the vaginal and cesarean delivery groups by hospital level of care. Approximately 89% of abdominally delivered gastroschisis patients were born at specialized facilities, whereas only 62% of vaginally delivered gastroschisis neonates were born at specialized centers (P < .001).
Before any further comparisons between vaginally and abdominally delivered infants with gastroschisis were made, we decided a priori to determine whether gastroschisis neonates actually differ from other malformation-free infants born in the same state with respect to fetal growth patterns in utero and neonatal survival. This analysis revealed that, on average, infants with gastroschisis were born 2 weeks earlier than normal singleton live births (mean ± standard deviation 36.9 ± 3.2 versus 38.9 ± 2.6; P < .001). They also weighed 813.9 g less than their anomaly-free counterparts (2,494.8 ± 594.3 versus 3,308.7 ± 615.4; P < .001).
The results comparing fetal growth indices between infants with gastroschisis and those who were anomaly-free are illustrated in Figure 2. Apart from being small for gestational age and postterm, infants with gastroschisis were disadvantaged in terms of all fetal growth parameters and neonatal death. Infants with gastroschisis were 7 and 4 times more likely to be of low birth weight and preterm, respectively. The occurrences of very low birth weight and very preterm were more than doubled in infants with gastroschisis. The widest disparity was with respect to neonatal survival, in which the survival probability was 10 times lower among infants with gastroschisis. The incidence of postterm was twice as high among anomaly-free babies whereas that of small for gestational age was 40% higher as compared with gastroschisis-affected neonates.
We then conducted analysis within the cohort of infants with gastroschisis stratified by mode of delivery. The mean birth weight and mean gestational age for the entire group of infants with gastroschisis was 2,494.8 ± 594.3 g and 36.9 ± 3.2 weeks, respectively. There were no significant differences in mean birth weight (2,468.8 ± 638.9 g versus 2,519 ± 549.9 g) and gestational age (36.9 ± 3.3 weeks versus 36.9 ± 3.0 weeks) between vaginally and abdominally delivered gastroschisis neonates, respectively. Other fetal growth parameters were similarly comparable (Table 2).
Overall, 18 neonatal deaths occurred among infants diagnosed with gastroschisis, yielding a neonatal mortality rate of 5.1%. Crude neonatal mortality rate was similar in both vaginal and cesarean delivery subcohorts (12/174 or 6.9% versus 6/180 or 3.3%). Adjusted estimates are displayed in Table 3. In the first model, adjusted hazard ratio for the association between mode of delivery and neonatal mortality was estimated, and the results show that vaginally and abdominally delivered infants with gastroschisis had comparable likelihoods for neonatal demise.
Next, we assessed 2 common neonatal morbidities (small for gestational age and preterm) in the pathway precipitating neonatal mortality among neonates with isolated gastroschisis. We sequentially added to model I the variable coding for preterm (model II) and small for gestational age (model III) separately. The objective was to determine whether both small for gestational age and preterm birth explained the higher-than-expected mortality observed among gastroschisis cases or whether only 1 of them was important in the causality pathway. Infants with gastroschisis who were born preterm were more than 3 times as likely to die in the neonatal period as their counterparts who were born at term (Table 3). However, infants with gastroschisis that also were small for gestational age had the same risk level for neonatal demise as those infants with gastroschisis that were appropriate for gestational age.
Contrary to our working hypothesis, we found comparable risk levels for neonatal mortality among patients diagnosed with isolated gastroschisis irrespective of mode of delivery. As a result, we had to reject our working hypothesis that cesarean delivery was protective of neonatal demise among these patients. Although few previous studies did suggest that patients with ventral wall defects delivered abdominally might have better outcomes,16–18 most of the reports in the literature found no quantifiable advantage of cesarean delivery.19–27 The latter findings are in agreement with the results in this study. The issue regarding the advantage of cesarean delivery is based on the belief that cesarean delivery protects abdominal viscera from birth trauma (tearing and avulsion injuries) and ischemia, prevents exposure of the uncovered viscera to the bacterial flora of the vagina, and removes the likelihood of birth dystocia caused by an exteriorized bowel.26
Although the main outcome of interest was neonatal mortality, we also examined differences in fetal growth indices between vaginally and abdominally delivered neonates. The 2 groups were similar in terms of the frequency of low and very low birth weight, preterm and very preterm, postterm and small for gestational age. We extended our analysis by exploring the morbidity pathway that could explain neonatal mortality controlling for mode of delivery. Our findings revealed that the subtype of growth perturbation in the fetus was more important than mode of delivery. The combination of gastroschisis and preterm in an infant more than tripled the risk for neonatal demise. In contrast, the combination of gastroschisis and small for gestational age did not alter the neonatal survival prognosis of these babies compared with age neonates who were appropriate for gestational age. This indicates that the component of low birth weight that predicts early mortality among gastroschisis babies is preterm birth and not small for gestational age. This finding may prove useful to care providers involved in the management of infants with gastroschisis.
Certain limitations must be borne in mind in interpreting our findings. The New York State Congenital Malformations Registry does not retrieve information on potential confounding characteristics, such as indications for cesarean delivery, prenatal or postnatal diagnosis, and whether referral was involved. We also had no information on defect size. Nevertheless, our investigation controlled for several potential confounding characteristics, including level of specialization of the health facility where delivery occurred.
A strength of this study is the large sample size on which the analysis was conducted. This series is the largest reported evaluating the influence of mode of delivery on neonatal mortality among infants with gastroschisis. In addition, the study is population-wide, based on surveillance reports of one of the largest birth defects registries in the country. Thus, reported risk estimates represent population averages, which are less biased by the effects of selective referrals. Nonetheless, we advise caution in interpreting our results because to be able to show that the difference in neonatal mortality in this study was significant, one would need a sample size of 640 infants with isolated gastroschisis in each arm. Hence, sample size limitation may have hindered us from detecting a significant difference in neonatal mortality between the 2 modes of delivery. In summary, we found comparable risk for neonatal mortality among vaginally and abdominally delivered patients with isolated gastroschisis. Preterm birth rather than small for gestational age was the pathway that precipitates early demise in these patients.
1. Salihu HM, Pierre-Louis BJ, Druschel CM, Kirby RS. Omphalocele and gastroschisis in the state of New York, 1992–1999. Birth Defects Res Part A Clin Mol Teratol 2003;67:630–6.
2. Salihu HM, Boos R, Schmidt W. Omphalocele and gastroschisis. J Obstet Gynaecol 2002;22:489–92.
3. Reid KP, Dickinson JE, Doherty DA. The epidemiologic incidence of congenital gastroschisis in Western Australia. Am J Obstet Gynecol 2003;189:764–8.
4. Rankin J, Dillon E, Wright C. Congenital anterior wall defects in the North of England, 1986–1996: occurrence and outcome. Prenat Diagn 1999;19:662–8.
5. Roeper PJ, Harris J, Lee G, Neutra R. Secular rates and correlates for gastroschisis in California (1968–1977). Teratology 1987;35:203–10.
6. Tan KH, Kilby MD, Whittle MJ, Beattie BR, Booth IW, Botting BJ. Congenital anterior abdominal wall defects in England and Wales 1987–93: retrospective analysis of OPCS data. BMJ 1996;313:903–6.
7. Hemmenki K, Saloniemi I, Kyronen P, Kekomaki M. Gastroschisis and omphalocele in Finland in the 1970s: prevalence at birth and its correlates. J Epidemiol Community Health 1982;36:289–93.
8. Martinez-Frias ML, Salvador J, Prieto L, Zaplana J. Epidemiological study of gastroschisis and omphalocele in Spain. Teratology 1984;29:337–82.
9. Penman DG, Fisher RM, Noblett HR, Soothill PW. Secular rates in incidence of gastroschisis in the south west of England in 1995. Br J Obstet Gynaecol 1998;105:328–31.
10. Segel SY, Marder SJ, Parry S, Macones GA. Fetal abdominal wall defects and mode of delivery: a systematic review. Obstet Gynecol 2001;98:867–73.
11. Druschel C, Sharpe-Stimac M, Cross P. Process of and problems in changing a birth defects registry reporting system. Teratology 2001;64(suppl 1):S30–6.
12. Honein M, Paulozzi L. Birth defects surveillance assessing the “gold standard”. Am J Public Health 1999;89:1238–40.
13. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol 1996;87:163–8.
14. Calzolari E, Bianchi F, Dolk H, Milan M. Omphalocele and gastroschisis in Europe: a survey of 3 million births, 1980–1990: EUROCAT Working Group. Am J Med Genet 1995;58:187–94.
15. Cox DR. Regression model and life tables. J R Statist Soc 1972;B34:187–220.
16. Fitzsimmons J, Nyberg DA, Cyr DR, Hatch E. Perinatal management of gastroschisis. Obstet Gynecol 1988;71:910–3.
17. Swift RI, Singh MP, Ziderman DA, Silverman M, Elder MA, Elder MG. A new regime in the management of gastroschisis. J Pediatr Surg 1992;27:61–3.
18. Sakala EP, Erhard LN, White JJ. Elective cesarean section improves outcomes of neonates with gastroschisis. Am J Obstet Gynecol 1993;169:1050–3.
19. Grundy H, Anderson RL, Filly RA, Callen P, Goldberg JD, Slotnick RN, et al. Gastroschisis: prenatal diagnosis and management. Fetal Ther 1987;2:144–7.
20. Bethel CA, Seashore JH, Touloukian RJ. Cesarean section does not improve outcome in gastroschisis. J Pediatr Surg 1989;24:1–3; discussion 3–4.
21. Mercer S, Mercer B, D'Alton ME, Soucy P. Gastroschisis: ultrasonographic diagnosis, perinatal embryology, surgical and obstetric treatment and outcomes. Can J Surg 1988;31:25–6.
22. Lewis DF, Towers CV, Garite TJ, Jackson DN, Nageotte MP, Major CA. Fetal gastroschisis and omphalocele: is cesarean section the best mode of delivery? Am J Obstet Gynecol 1990;163:773–5.
23. Moretti M, Khoury A, Rodriquez J, Lobe T, Shaver D, Sibai B. The effect of mode of delivery on the perinatal outcome in fetuses with abdominal wall defects. Am J Obstet Gynecol 1990;163:833–8.
24. Sipes SL, Weiner CP, Sipes DR 2nd, Grant SS, Williamson RA. Gastroschisis and omphalocele: does either antenatal diagnosis or route of delivery make a difference in perinatal outcome? Obstet Gynecol 1990;76:195–9.
25. Adra AM, Landy HJ, Nahmias J, Gomez-Marin O. The fetus with gastroschisis: impact of route of delivery and prenatal ultrasonography. Am J Obstet Gynecol 1996;174:540–6.
26. How HY, Harris BJ, Pietrantoni M, Evans JC, Dutton S, Khoury J, et al. Is vaginal delivery preferable to elective cesarean delivery in fetuses with a known ventral wall defect? Am J Obstet Gynecol 2000;182:1527–34.
27. Segel SY, Marder SJ, Parry S, Macones GA. Fetal abdominal wall defects and mode of delivery: a systematic review. Obstet Gynecol 2001;98:867–73.