CHRISMED Journal of Health and Research

: 2020  |  Volume : 7  |  Issue : 1  |  Page : 51--55

Single institution experience with esophageal atresia over 9 years with a review of literature: Where do we stand?

Nandini Kaul Bedi1, Alka Gupta Grewal2, William Bhatti1,  
1 Department of Paediatric Surgery, Christian Medical College and Hospital, Ludhiana, Punjab, India
2 Department of Paediatric Surgery, Dayanand Medical College and Hospital, Ludhiana, Punjab, India

Correspondence Address:
Nandini Kaul Bedi
Department of Paediatric Surgery, Christian Medical College Hospital, Brown Road, Ludhiana . 141 008, Punjab


Aim: We have conducted a study of the clinical profile of esophageal atresia (OA) at our institution over a period of 9 years with an aim toward improving our results. We have attempted to find the significant indicators of mortality to be able to focus on improving the survival of these patients in future. To achieve this end, we applied the Waterston's prognostic classification and the Spitz' prognostic criteria to our patients and studied the effects of each variable independently as a cause of mortality. We have also compared the two prognostic criteria to see, which has more significance for us at our institution. Materials and Methods: Over a period of 9 years, from 2006 to 2014, 73 neonates were treated for OA at our institute. There were 27 mortalities. We analyzed our results using the Waterston and the Spitz criteria and studied each independent variable within the two prognostic classifications for the more highly significant cause of mortality. The variables considered as significant indicators of mortality were low birth weight (LBW), pneumonia, associated congenital heart disease (CHD), other associated anomalies, and very LBW. Results: The Waterston prognostic classification was very highly significant (P < 0.0005) as an indicator of prognosis in this study. The two most significant variables affecting survival were LBW, as described by Waterston (P = 0.005) and the association of major CHD (P = 0.006). We had 63% survival and none below the birth weight of 1.8 kg. Conclusion: The study clearly points out that LBW is a highly significant indicator of prognosis followed by the association of major CHD. Waterston's risk classification continues to be a highly significant indicator of prognosis. Further studies are required to assess the combined effects of more than one variable on babies with OA.

How to cite this article:
Bedi NK, Grewal AG, Bhatti W. Single institution experience with esophageal atresia over 9 years with a review of literature: Where do we stand?.CHRISMED J Health Res 2020;7:51-55

How to cite this URL:
Bedi NK, Grewal AG, Bhatti W. Single institution experience with esophageal atresia over 9 years with a review of literature: Where do we stand?. CHRISMED J Health Res [serial online] 2020 [cited 2020 Aug 8 ];7:51-55
Available from:

Full Text


Before 1939, there were no survivors of the anomaly of oesophageal atresia (OA) with or without (±) tracheo-esophageal fistula (TOF). Cameron Haight of Michigan, in 1939, is credited with the first survivor with a primary repair of the esophagus.[1] He used a left thoracic approach and a single layer anastomosis. In the same year, Levin[2] of St Paul and Ladd[3] of Boston also reported survivors of the anomaly but without a primary repair. They did a gastrostomy and a delayed ligation of the TOF. Haight in 1943 revised his technique to a right thoracic approach. He also classified the anomaly,[4] but the more condensed classification of Gross[5] is universally followed today.

With time came the understanding of the indicators of prognosis and Waterston et al. in 1962[6] assigned these babies to three risk groups based on the presence of other congenital anomalies and pneumonia. Overall survival improved from 36% in the 1950s[1] to more than 90% in the present era.[7] A new risk classification was proposed by Spitz et 1994[8] based on very low birth weight (VLBW) and the presence of congenital cardiovascular anomalies. Niramis et al. in 2013[7] found, on comparing, the Spitz criteria to be more useful in predicting the prognosis of their patients than the Waterston's criteria. This clearly reflects an improvement in the all-round care of these patients.

OA ± TOF with a wide gap remains a challenge down the decades. Ein et al.,[9] Ein and Shandling[10] from Canada; Al-Shanafey and Harvey[11] from Australia and Holland et al.[12] from UK have reported their management of long gap OA mostly without TOF. They report the changing trend of management from esophagostomy with gastrostomy followed by delayed replacement of esophagus, mainly with the gastric tube or gastric pull up as against retaining the native esophagus with a delayed primary repair. Retaining the native esophagus has shown the best results and is increasingly becoming the procedure of choice in the developed countries. Glasson et al.[13] have reported this change in the surgical approach and stressed the need for a multi-center database development for further refinement in the management of these patients.

We are presenting the results of our center in a retrospective study of 9 years from 2006 to 2014, wherein 73 neonates presented with the diagnosis of OA ± TOF. We have attempted to understand the leading cause of mortality in these babies. The five main associated variables studied are low birth weight (LBW), pneumonia, major congenital heart disease (CHD), other associated anomalies, and VLBW. To study each variable as the main cause of mortality, we classified our patients according to the Waterston's and the Spitz' prognostic classification and compared the significance of each in our institutional setup. Each variable was then studied as the leading cause of mortality.

 Materials and Methods

Over a period of 9 years from 2006 to 2014, all patients of OA ± TOF that presented to our hospital were included in this retrospective study. After the Institutional Review Board and ethical clearance, informed consent was obtained from guardians of all patients that fulfilled the criteria of inclusion in this study. Patient records were taken out from the institutional database, and their management and results were studied and analyzed. A total of 75 patients presented to our institute with a history suggestive of OA ± TOF, of which 73 were neonates. Two patients were referred for secondary procedures and are excluded from the study. Of 73 patients, 66 (90.4%) had Type C atresia (Gross' classification), and 7 (9.5%) patients had pure OA, Type A.

Only ten mothers had an antenatal diagnosis of polyhydramnios, which could not be taken into account as all mothers did not undergo an antenatal sonogram. None were diagnosed to have OA antenatally. Of the 66 babies with Type C OA, 14 neonates (19.1%) were preterm and 52 neonates (71.2%) had LBW. Birth weight ranged from 1.25 to 3.6 kg. Out of the seven babies with pure OA, two were preterm, and six had LBW. The birth weight ranged from 1.7 to 2.75 kg. Associated anomalies with organ system distribution are listed in [Table 1].{Table 1}

There were six babies who died before surgery. All were OA Type C, and three were preterm. One baby weighed 2.9 kg; all others were LBW, ranging from 1.25 to 1.9 kg. Two babies had major congenital cardiac anomalies, one had hyaline membrane disease; five babies had severe, and one had moderate pneumonia. They did not respond to resuscitative measures.

Primary repair was attempted in the remaining 60 neonates with Type C-OA with TOF. In 52 patients, the right thoracotomy, the extrapleural approach was used, and resection of TOF with the primary repair was done in 51 patients. One patient had a wide gap which could not be bridged; hence, an esophagostomy and gastrostomy were done, he later succumbed to sepsis. Thoracoscopic repair was attempted in eight patients and three were converted to open repair. One patient had a sudden cardiac arrest on the table and could not be revived. In all, 43 patients did well after the primary repair, and 17 patients died. Of these patients who died, nine were term, and eight were preterm babies, all were LBW, ranging from 1.3 to 2.4 kg (median weight 1.65 kg). Eight babies did not have any associated anomalies, but seven had pneumonia leading to septicemia. Four had major cardiac anomalies, two had the VACTERL group of anomalies, and one had Down's syndrome.

Seven babies with pure OA Type A are included in this study. Four babies who were term but LBW ranging from 1.9 to 2.1 kg underwent esophagostomy and gastrostomy. All four had associated cardiac anomalies. Two of them succumbed to sepsis in the postoperative period. Two did well and went home. One of whom was lost to follow-up and the other did well and came back at 8 months to undergo a gastric pull-up. He is well and in follow-up. A primary repair was done in the earliest baby, the details of which are not recorded, but this baby died from sepsis. Another preterm, 1.7 kg baby with associated cardiac anomaly, underwent a primary gastric pull-up with a feeding jejunostomy; he too died from sepsis. The most recent patient was a 2.75 kg term baby who underwent a primary esophageal repair[14] and is doing well.


Seventy-three neonates were treated in our hospital for OA ± TOF over 9 years. Out of these seven had Type A and 66 had Type C-OA. We had 27 mortalities, of which 4 were Type A, and 23 were Type C-OA. Six of the babies with OA Type C died before surgery. In an attempt to analyze our result, we classified our patients according to the Waterston's prognostic classification as well as the currently used Spitz' prognostic criteria. [Table 2] shows the distribution of our patients according to each classification as well as the mortality and survival of each group. According to Waterston's prognostic classification, most of the patients fell in Group B (n 38) and Group C (n 32). Their survival was 94.7% and 21.8%, respectively. Very few were classed as Group A (n 3), and survival was 100%. We then classified our patients according to the Spitz' criteria, and the majority fell into Grade I (n 54) with 74% survival, 17 patients were in Grade II with 35.2% survival; only two patients fell into Grade III with no survivors. Waterston and Spitz criteria were compared using the area under the curve (AUC); values of P < 0.05 were considered statistically significant. Categorical variables are presented in numbers and percentages. Odds ratio with 95% confidence interval was calculated for selected variables, and their significance tested. Univariate and multivariate logistic regression was used to assess the association between mortality and several potential risk factors. The data were analyzed using the SPSS version 21.0 (IBM, Chicago, USA).{Table 2}

Although the Spitz criteria was a significant indicator (P = 0.003) of prognosis, the Waterston classification was a more highly significant indicator (P < 0.0005) of prognosis and mortality [Table 2]. The most significant cause of mortality was LBW <2.5 kg (P = 0.005) followed by associated major CHD (P = 0.006). VLBW, pneumonia, and other associated anomalies were not statistically significant indicators of mortality [Table 3]a. In a multivariate logistic regression analysis [Table 3]b, LBW <2.5 kg is the single, highest, independent risk factor for mortality. Associated major CHD was also a significant indicator of mortality (P = 0.006), but the number of patients was too few to be conclusive. VLBW was also significant, but again, there were very few patients. Pneumonia and other associated anomalies were not sufficiently significant indicators of mortality [Table 3]a.{Table 3}


According to our analysis, the Waterston classification had a much higher predictive value with an AUC of 0.889 and a standard error of 0.042, as compared to Spitz's classification, which has an AUC of 0.680 and standard error of 0.069. In our study, 52 (71.2%) patients had LBW, 43 (58.9%) had pneumonia, 36 (49.3%) had major associated anomalies, only 14 (19.1%) had major CHD and 4 (5.4%) babies had VLBW (<1.5 kg). Statistically, LBW (n = 52) was the single most significant factor predicting prognosis, followed closely by associated major CHD (n = 14). Pneumonia is not a significant predictor of mortality. Hence, we look toward LBW and its effects, causing poor outcomes in our patients with OA. Major CHD is also a significant factor, but second to LBW as the number of patients having CHD of significance were not many (n = 14).

Looking at the contradictory fact that though both prognostic classifications are significant in our study, Waterston's classification of risk factors which includes LBW, with a demarcation at 1.8 kg, moderate-to-severe pneumonia, and association of moderate-to-severe congenital anomalies has a much higher significance and predictive value in our study. Spitz' prognostic criterion considers birth weight <1.5 kg (VLBW) as the demarcating line and association of major CHD as an equally significant factor. The presence of both places the patient in the highest risk, presence of either of the two, in intermediate-risk and the absence of both puts the patient in the lowest risk group. Pneumonia is not a risk factor in this classification. In our study, pneumonia was not a significant predictor of mortality, and yet the Waterston prognostic classification is a very highly significant predictor of prognosis. Clearly, demarcation at a higher birth weight is an important prognostic factor. We have higher survival above the birth weight of 1.8 kg (79.3%) than the above birth weight of 1.5 kg (66.66%).

Pneumonia is not a statistically proven risk factor, but can we completely ignore it? Out of the 27 mortalities, 19 babies suffered from pneumonia (70.3%). One baby had normal birth weight with associated major CHD, complicated with pneumonia. The remaining 18 babies were LBW, and 4 had associated CHD while 14 had LBW complicated with pneumonia alone.

It is probably time to rethink about the outcome and the factors affecting that outcome in babies with OA. Segregation of the babies who are at higher risk as compared to others and focusing on their specific problems is required. Maternal health programs need to be upgraded and implemented. Studies similar to ours need to be done in different parts of our country to collect data on a larger scale to be able to come to a definite conclusion and improve the outcome of our babies with OA.


In a retrospective study conducted at our institution over a 9-year period, we studied 73 patients who presented with OA ± TOF, of these 27 died (36.98%) and in an attempt to find out the leading cause of mortality we classified them according to the Waterston's prognostic classification as well as the Spitz' prognostic criteria and compared our results. We found that the Waterston's classification was a very highly significant predictor of prognosis (P < 0.0005) as compared to the Spitz' criteria. We also studied the associated conditions individually as the leading cause of mortality and found LBW to be the most significant cause (P = 0.005), followed by major CHD (P = 0.006). Although pneumonia, VLBW, and other associated anomalies were not sufficiently significant indicators of mortality, their numbers were few, and as such, a larger study, possibly over a longer time span, is required to comment definitely.


We are extremely grateful to Mrs. Bhavna Garg, BSc (Hon), Chartered Financial Analyst, who not only helped us with the statistical analysis of our study but also took the time and patience to answer all our queries, big and small.

We are also very grateful to Mrs Darshana Devi, Social Worker, who spent endless hours contacting guardians of our patients to help in completing this study.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Manning PB, Morgan RA, Coran AG, Wesley JR, Polley TZ Jr., Behrendt DM, et al. Fifty years' experience with esophageal atresia and tracheoesophageal fistula. Beginning with Cameron Haight's first operation in 1935. Ann Surg 1986;204:446-53.
2Levin NL. Congenital atresia of the esophagus with tracheoesophageal fistula: Report of successful extra pleural ligation of fistulous communication and cervical esophagotomy. J Thorac Surg 1941;10:648-57.
3Ladd WE. The surgical treatment of esophageal atresia and tracheoesophageal fistulas. N Engl J Med 1944;230:625-37.
4Haight C. Congenital esophageal atresia and tracheoesophageal fistula. In: Mustard WT, Ravitch MM, Snyder WH, Welch KJ, Benson CD, editors. Pediatric Surgery. 2nd ed. Chicago: Year Book Medical Publishers; 1969. p. 357-82.
5Grosfeld JL, O'Neill JA Jr., Fonkalsrud EW, Coran AG, editors. Pediatric Surgery. 6th ed., Ch. 67. Philadelphia: Mosby, Elsevier; 2006. p. 1053.
6Waterston DJ, Carter RE, Aberdeen E. Oesophageal atresia: Tracheo-oesophageal fistula. A study of survival in 218 infants. Lancet 1962;1:819-22.
7Niramis R, Tangkhabuanbut P, Anuntkosol M, Buranakitjaroen V, Tongsin A, Mahatharadol V. Clinical outcomes of esophageal atresia: Comparison between the Waterston and the Spitz classifications. Ann Acad Med Singapore 2013;42:297-300.
8Spitz L, Kiely EM, Morecroft JA, Drake DP. Oesophageal atresia: At-risk groups for the 1990s. J Pediatr Surg 1994;29:723-5.
9Ein SH, Shandling B, Heiss K. Pure esophageal atresia: Outlook in the 1990s. J Pediatr Surg 1993;28:1147-50.
10Ein SH, Shandling B. Pure esophageal atresia: A 50-year review. J Pediatr Surg 1994;29:1208-11.
11Al-Shanafey S, Harvey J. Long gap esophageal atresia: An Australian experience. J Pediatr Surg 2008;43:597-601.
12Holland AJ, Ron O, Pierro A, Drake D, Curry JI, Kiely EM, et al. Surgical outcomes of esophageal atresia without fistula for 24 years at a single institution. J Pediatr Surg 2009;44:1928-32.
13Orford J, Cass DT, Glasson MJ. Advances in the treatment of oesophageal atresia over three decades: The 1970s and the 1990s. Pediatr Surg Int 2004;20:402-7.
14Bedi NK, Grewal AG, Rathore S, George U. A new approach for the management of esophageal atresia without tracheo-esophageal fistula. J Indian Assoc Pediatr Surg 2016;21:144-6.