Introduction

Intestinal atresias are one of the most common causes of intestinal obstruction in the neonate with an incidence of 1 in 5,000 newborns [1, 2]. The prognosis for this condition has improved significantly since the early 1970s [2–4]. This improvement has coincided with advances in diagnostic modalities, surgical technique, neonatal intensive care and parenteral nutrition. We present our results in this condition comparing the incidence of specific clinical features and surgical outcomes in the different subsets of atresias.

Materials and methods

A retrospective chart review of all infants undergoing surgical correction of intestinal atresia over a 25-year period (1983–2008) was conducted at our tertiary care institution after obtaining institutional review board approval.

Duodenal atresia diagnosis was based on identifying an antenatal double bubble on ultrasound or on plain X-ray demonstration of the double bubble sign postnatally. The stomach was decompressed with a replogle tube, while cardiac defects were excluded by an echocardiogram. After adequate evaluation of other associated congenital anomalies, surgical repair was planned. All repairs of this anomaly were performed through a right upper quadrant transverse laparotomy. Location of ampulla of vater was confirmed by gentle compression of the gallbladder prior to repair. We confirmed the absence of concomitant distal atresias by placing a feeding tube in the duodenotomy and confirming free passage of irrigated fluid to the cecum. A windsock deformity was excluded by ensuring that the nasogastric tube could be manipulated through the proximal duodenotomy. The management of duodenal atresia at our institution involved a duodeno-duodenostomy of the side-to-side type, a diamond-shaped anastomosis, a conservative web excision, and in the more distal types of duodenal atresias a duodeno-jejunostomy. Oral intake was permitted once the nasogastric aspirate had decreased and bowel function returned. No trans-anastomotic tubes were used in any of our patients.

For the more common jejuno-ileal atresias, we always performed a water-soluble contrast enema to exclude any concomitant colonic atresia after establishing the diagnosis of small bowel obstruction by supine and decubitus views of the abdomen. Atresias were classified based on the Grosfeld modification of the intestinal atresia classification [5]. In the type I (web), II (fibrous cord) and IIIA (mesenteric defect) atresias, an end-to-end anastomosis was the most commonly performed procedure with or without tapering of the dilated proximal bowel. Bowel length conservation was more important in the more complicated type IIIB (apple peel) and type IV (multiple) atresias. We anastomosed the dilated proximal bowel to the distal apple peel based on the often tenuous ileo-colic branch. Anastomosis over inert silastic tubing was used in multiple atresias. For the rare colonic atresia, a primary anastomosis was performed for right-sided lesions and a colostomy was performed for left-sided lesions (splenic flexure and beyond) during the initial operation. Infants with gastroschisis were excluded from the evaluation of differences in birth weight.

Oral intake was allowed once bowel function returned. All neonates were supported with parenteral nutrition until goal calories were achieved by oral intake. The feeding regimens remained the same over the study period but the formulas available for infants with short bowel were changed. We started feeds slowly in these premature infants and increased the rate first and increased osmolarity (concentration) of feeds [if needed] only after goal rates had been achieved and tolerated. Most neonates were gavage fed until they were mature enough to swallow.

Data collection included prenatal diagnosis, demographics, birth weight, gestational age, presenting symptoms, associated abnormalities, incidence of abdominal wall defects, radiological findings, surgical intervention, morbidity, surgical re-intervention, time to establishment of full oral intake, length of hospital stay and mortality. The Fisher test was used to test the significance of differences in categorical data, while the student’s t test and ANOVA were used to analyze the significance of differences between continuous variables.

Results

A total of 130 atresias were repaired during the study period. Table 1 illustrates the various clinical features of the three groups. The male to female ratios in duodenal atresia were 1:1 (29 male, 30 female), in jejuno-ileal atresia 1.17:1 (34 male, 29 female), and in colonic atresia 0.6:1 (3 male, 5 female). Antenatal diagnosis was established in 27 of 59 babies (46%) with duodenal atresia, 26 of 63 (41%) with jejuno-ileal atresia and 1 of 8 (12.5%) with colonic atresias (p > 0.05). The prenatal ultrasound findings included ultrasound “double bubble” in duodenal atresia, dilated echogenic bowel loops in jejuno-ileal and colonic atresias, and polyhydramnios in a proportion of both proximal and distal atresias. The mean gestational ages were 36 weeks in duodenal atresia, 37 weeks in jejuno-ileal atresia, and 37 weeks in colonic atresia. The mean birth weights (excluding those with gastroschisis) were 2,380.5 g (SD 988) in duodenal atresia, 2,814 g (SD 755) in jejuno-ileal atresia, and 3,153 g (SD 527) in colonic atresia (p = 0.011). In the jejuno-ileal atresia group, there were 14 type I, 14 type II, 16 type IIIA, 9 type IIIB and 10 type IV atresias. In the colonic atresia group, four had a type I lesion and four had a type II defect. Associated anomalies were seen in 41 (76%) patients with duodenal atresia, 32 (52%) with jejuno-ileal atresia, and 3 (38%) with colonic atresias (p > 0.05). Down’s syndrome was seen in 18 (31%) of the patients with duodenal atresia. In this series, gastroschisis was observed in ten infants with colonic and jejuno-ileal atresia (14%). Seven of these infants had jejuno-ileal atresia, one had a colonic atresia, and two had atresias in both the colon and the small intestine. Table 2 illustrates the incidence of different congenital anomalies among the three groups.

Table 1 Clinical characteristics of neonates born with intestinal atresia
Full size table
Table 2 Incidence of associated anomalies in neonates born with intestinal atresia
Full size table

In the duodenal atresia group, bilious vomiting was the most frequent mode of presentation, while in the colonic atresia group abdominal distension was the most frequent presentation with vomiting being a late feature. Other presentations in duodenal atresia included loss of weight or failure to gain weight (8%), abdominal distension secondary to gastric dilatation (14%), aspiration (3%), and delayed passage of meconium (3%). The presentation in the jejuno-ileal atresia group was more variable and was dependent on its location vis-à-vis the duodenum. Of the 30 patients with more proximal atresias, a majority presented with bilious emesis while abdominal distension was the more common presentation in the 33 patients with distal small bowel atresia. Duodenal atresia was diagnosed clinically and confirmed by plain X-ray. An upper GI was performed rarely and only in instances where the child was older and malrotation with volvulus was a possibility. All patients with jejuno-ileal and colonic atresias underwent water-soluble contrast enemas to define colonic anatomy. Of the colonic atresias, three were in the right colon, two were in the transverse colon, and three were in the sigmoid colon.

In neonates with duodenal atresia, duodeno-duodenostomy was the most frequently performed procedure and was performed in 50 patients. Duodenojejunostomy was done in four patients and a conservative web excision, with preservation of the medial wall, was performed in four others. Resection and anastomosis was performed in one patient. Ten infants who had duodenal atresia needed re-hospitalization secondary to complications related to the initial operation. Eight of these patients were re-operated upon, while one underwent radiology guided balloon dilatation for anastomotic stenosis and the other had three admissions for adhesive sub acute bowel obstructions that were conservatively managed. Of the eight re-operations, six (10%) needed surgery (lysis of adhesions) due to adhesive bowel obstruction, one needed a re-exploration for anastomotic stenosis, and the last patient needed re-exploration for a missed distal atresia. In neonates with jejuno-ileal atresia, 52 underwent resection and end-to-end anastomosis, five patients underwent a Bishop-Koop procedure, and five had a stoma created. One patient with gastroschisis and a known atresia developed total small bowel necrosis and underwent an exploration and closure only. Primary tapering (at the initial operation) was performed in four patients (6%). Nine patients (14%) needed re-exploration for adhesive bowel obstruction and one patient needed re-intervention for an anastomotic leak. Intestinal dilatation developed in 7 of 63 patients (11%). A surgical re-exploration and tapering was performed in five patients at 16, 18, 22, 28 and 90 days after the initial atresia surgery due to feeding intolerance and functional obstruction on oral contrast studies. Two patients, one with a type IIIB and the other with a type IV atresia and bowel dilatation underwent a Bianchi [6] and STEP [7] procedure with good results. The Bianchi procedure was done at 1 year and 9 months of age and the STEP (serial transverse enteroplasty) procedure at 3 years of age. In the colonic atresia group, five underwent resection and anastomosis and three underwent a Bishop-Koop procedure (end-to-side anastomosis with the end of the distal bowel being exteriorized as a chimney). One patient needed revision of the end-to-side colonic anastomosis at the time of the Bishop-Koop closure due to stenosis.

Time to full oral intake was less than 2 weeks in nearly 45% of neonates with duodenal atresia. In the jejuno-ileal atresia group, the mean duration on parenteral nutrition was 20 days (SD 17.8, third quartile = 22.75). Anyone on nutritive support for longer than 40 days was considered an outlier based on the Tukey’s test of variance. There were 11 outliers based on the Tukey test. Six of them had gastroschisis, one had a choledochal cyst, one had associated cardiac anomalies, one had a high output stoma, and one neonate had a complex syndrome (syndrome of Kamouraska: an autosomal recessive condition associated with peripheral neuropathy, psychomotor retardation, chronic diarrhea and increased long chain fatty acids). In the colonic atresia group, average duration of parenteral nutrition requirement was 78 days in the group with gastroschisis and 11 days in the group without.

Infants with duodenal atresia had an excellent prognosis with no mortality in our series. The mean duration of hospitalization was 32.6 days (SD = 34.6 days) and the mean surgical follow-up was 12.5 months (SD = 15.8). Trisomy 21 was not associated with poor prognosis. Associated anomalies (p = 0.0015), prematurity (p = 0.0119) and low birth weight (p = 0.0110) increase the duration of hospitalization without affecting survival. In the jejuno-ileal atresia group, the average hospital stay was 41 days (8–322 days; 95% CI 30.3–63.1). Outliers determined based on the Tukey test were patients with a stay greater than 98 days. There were four patients with significantly longer inpatient stay. Two of these patients had gastroschisis and two others had apple peel atresias (type IIIB). The overall mortality rate in the jejuno-ileal atresia group was 11%. Most patients who died had associated anomalies (p = 0.017) or type IIIB/IV atresias (p = 0.007). There were three in-hospital deaths due to severe atresias and multiple anomalies. The anomalies included porencephaly (1), absent septum pellucidum and gastroschisis (1), and structural cardiac defect (1). Three infants died due to complications linked to parenteral nutrition-related cholestasis and sepsis. These were late deaths occurring at 1.6, 2 and 3 years after initial surgery. In this group too, two patients had associated anomalies including a choledochal cyst (1) and the syndrome of Kamouraska (1). One patient with type IIIA jejuno-ileal atresia died 2.5 months after surgery due to sepsis from an undetermined source after an uneventful recovery from surgery. In the colonic atresia group, the average duration of hospital stay was 98 days in those with abdominal wall defects and 10 days in the group without. There was no mortality in this group.

Discussion

Nearly one-third of all congenital bowel obstructions are due to intestinal atresia [1, 2]. The etiology of this condition is thought to be secondary to incomplete vacuolization of the solid primitive duodenum [8] and secondary to vascular or mechanical accidents in the case of the jejunum/ileum and colon [9]. Our accuracy in detecting small bowel atresias is around 45% and this compares favorably with another series with similarly collected data and accuracy rates of between 23 and 31% [12]. In our experience, all signs of intestinal obstruction (stomach “double bubble”, dilated echogenic loops >7 mm in internal diameter and free peritoneal fluid) are most accurately detected from the late second trimester onward. This coincides with the process of the return of the intestine back into the coelomic cavity from the physiological herniation into the extraembryonic coelom (10–12 weeks of intra-uterine life). The critical sequence of events at this point of time involves intestinal rotation and fixation and hence lends credence to our hypothesis that the association between malrotation and atresia may be due to an abnormality in this physiological process. Similar results have been described by others [10, 11]. Further, our study reaffirms that proximal intestinal obstruction appears to be detected more easily than obstruction further distally in the intestinal tract. This is explained by the ability of the colon to absorb fluid, thereby preventing significant intestinal dilatation in distal obstruction [13].

There was a significant difference in birth weight of nearly 400 g between atresias at different levels in the intestine. This could not be explained on the basis of maternal factors, prematurity or growth restriction because of associated congenital anomalies, which were not significantly different between groups. Our study was the first to associate the length of mucosal surface to higher birth weight [14]. Inter-related work by Blakelock et al. demonstrated that term babies born with gastroschisis are significantly growth retarded compared with premature babies born with gastroschisis. Term babies born with a proximal intestinal atresia in gastroschisis were also growth retarded, which suggested that in late gestation, the normal growth is dependent on a normally functioning gastrointestinal tract that allows exposure of the proximal intestinal mucosa to ingested amniotic fluid [15]. Others have also shown greater growth retardation in neonates with jejunal as compared to ileal atresia [16, 17]. We hypothesize, based on our clinical observations to date, of a possible entero–endocrine axis that contributes to fetal weight gain beyond the third trimester. We suggest that the amniotic fluid glucose and insulinogenic amino acids (such as ornithine and arginine found in high concentrations in the amniotic fluid) play an important role in regulating fetal insulin levels through modulation of incretin peptides GLP-1 (glucagon-like peptide) and GIP (glucose-dependent insulinogenic peptide). It is well established that GLP is secreted by K cells located in the proximal duodenum, while the L-cells that secrete GIP are located in the distal small intestine and colon [18]. We suspect that because of the ability of these specialized mucosal cells to increase beta cell mass and promote insulin release, they play an important role in the fetus in promoting third trimester weight gain. Our observation of distal atresias having greater weights than proximal atresias supports this anatomic distribution of incretin peptides in the intestine.

Associated anomalies were observed in nearly 75% of infants with duodenal atresia. This is higher than the incidence of associated anomalies noted in about 50% of newborns analyzed in another large series published recently [19]. Nearly one-third of these neonates have Down’s syndrome, but their prognosis is no worse than those without the condition. A similar observation has been reported by others [20]. Further research is needed to explain the overall high incidence of associated anomalies (52%) in the jejuno-ileal atresia group. We suspect that based on earlier work done in this geographic area, there is a higher incidence of genetic transmission of intestinal atresia [21].

In duodenal atresias, duodeno-duodenostomy is the operation of choice and avoids the complications of duodeno-jejunostomy including delayed anastomotic function and blind loop syndrome that have been reported from our center earlier [22]. Based on our experience, tapering should be performed at the initial exploration if there is significant luminal disparity between the two ends, particularly in the proximal jejunal atresias. Other options include resection of the dilated bowel up to normal caliber and perform anastomosis at that level. This would be applicable in infants with adequate bowel length. The Bishop-Koop procedure is an option for more distal atresias with a significant luminal disparity and helps conserve mucosal absorptive surface. Adhesive bowel obstruction remains a significant cause of morbidity after surgery for intestinal atresia. In this series, around 12% of the neonates needed re-operative surgery for bowel obstruction after atresia correction. These figures are similar to another large series [10]. The clinical presentation of bowel obstruction could theoretically occur at any point, but interestingly in our series of jejuno-ileal atresia all presented within the first year after atresia correction. Bowel dilatation may be a blessing in disguise for infants with short bowel syndrome allowing for bowel lengthening procedures such as the STEP or Bianchi.

There was no significant difference between atresia subgroups in the time to cessation of parenteral nutrition. Intestinal atresia associated with gastroschisis caused a significant prolongation in the duration of parenteral nutrition requirement and hospital stay. Length of hospital stay was also significantly prolonged in patients with intestinal atresia and prematurity or in the presence of significant associated anomalies. The mortality in our series was observed only in the jejuno-ileal atresia group (11%). This is similar to data from other series [1, 2, 10]. Improvement in neonatal care and enteral and parenteral nutrition technology has significantly improved the prognosis in these neonates.