Pediatr Surg Int (2004) 20: 757–761 DOI 10.1007/s00383-004-1293-5

O R I GI N A L A R T IC L E

Simmi K. Ratan Æ Alok Varshney Æ Sumita Mullick Naresh Chand Saxena Æ Sunil Kakkar Æ Punita K. Sodhi

Evaluation of neonates with esophageal atresia using chest CT scan

Accepted: 4 September 2004 / Published online: 27 October 2004 Ó Springer-Verlag 2004

Abstract Recent literature stresses various factors that influence outcome in neonates with esophageal atresia (EA), including associated congenital anomalies, ventilator dependency, and gap length. Although the existing modalities can fairly accurately assess the first two factors, a comparable modality for assessing the interpouch gap in a natural unstretched position preoperatively is lacking. A technique of direct sagittal computed tomography (CT) scanning of the chest, initially employed by Tam et al. in 1987, was found to provide fairly accurate information on the level of origin of the fistula and the interpouch distance, with pouches in natural, unstretched positions, in two neonates with EA and distal tracheoesophageal fistula (TEF). However, no further studies were reported using this technique. We studied 10 neonates with EA and distal TEF, the most common variant, using this technique. The addition of axial cuts was done in the event of a short upper pouch or suboptimally-clear anatomical details in sagittal chest CT films. The preoperative information in our study subjects was found to corroborate well with operative findings except in the one patient with laryngotracheoesophageal cleft (LTEC). Prior information of a left bronchoesophageal fistula in two of the neonates with dextrocardia greatly helped us in saving operative time and difficulties and postoperative hassles. The chest CT scan proved to be a noninvasive and quick means for investigating neonates with the S. K. Ratan (&) E-13/13 Khirki Extension, Malviya Nagar, New Delhi PIN 110017, India E-mail: [email protected] Tel.: +91-11-6239889 S. K. Ratan Æ S. Mullick Department of Paediatric Surgery, Safdarjang Hospital, New Delhi, India A. Varshney Æ N. C. Saxena Æ S. Kakkar Department of Radiodiagnosis, Safdarjang Hospital, New Delhi, India P. K. Sodhi Department of Ophthalmology, Safdarjang Hospital, New Delhi, India

most common variant of EA. CT scanning could be extended for routine preoperative evaluation of neonates with EA to prepare for intraoperative problems, besides helping prognostication. Keywords Esophageal atresia Æ Neonates Æ CT scan Æ Investigation Æ Tracheoesophageal fistula

Introduction Each case of esophageal atresia (EA) has to be assessed individually based on a number of factors, including type of atresia, weight of the neonate, associated anomalies, origin of the fistula, and interpouch distance. The two primary factors that influence outcome in these patients are the primary esophageal pathology and the state of the patient [1]. The existing prognostic classifications, by and large, concentrate on the latter [3, 5, 7– 9]. Some authors have classified neonates with EA into three categories based on the gap between the upper and the lower esophageal ends [6, 10]. They found that the interpouch gap is highly fairly predictive of mortality and, especially, long-term morbidity (in terms of leak and strictures), irrespective of the neonate’s size. According to them, this classification could be immensely useful for the parents and the health planners. We have found that the interpouch gap dictated not only the mortality and long-term morbidity in these patients but also guided the preoperative and postoperative strategies, including the need for postoperative ventilatory assistance. With limited facilities for neonatal ventilatory assistance, we needed an investigative modality that could fairly predict the gap between the two esophageal ends in their natural position preoperatively. Unfortunately, none of the currently used modalities for investigating these tiny patients was found to furnish all the desired information regarding the exact level of the esophageal pouches, level of the fistula, and interpouch distance.

758

In 1987, Tam et al. introduced a technique of employing direct sagittal computed tomography (CT) scanning of the chest to evaluate these parameters in two neonates with EA with distal tracheoesophageal fistula (TEF) [10]. With depiction of the trachea and esophagus in their natural positions and the provision to measure interpouch distance, they found this technique to be extremely useful. Later, they modified the same technique to diagnose an H-type fistula in another neonate [11]. However, no further studies were reported using this technique. The present study was undertaken to analyze the diagnostic utility of the modality of chest CT scanning in neonates with EA.

Material and methods During a period of 6 months from December 1998 to May 1999, 15 neonates were admitted with esophageal atresia to the pediatric surgical unit of Safdarjang Hospital. These included referrals from the inhospital labor room/nursery and other hospitals of northern and western India. All the neonates were full-term deliveries with birth weights ranging between 2.0 and 2.8 kg. The reporting time varied from 1 h for the inhospital-delivered neonates to 72 h for the referrals from distant places. The referrals were accompanied by plain chest radiographs of the neonates in posteroanterior and/or lateral position with a coiled nasogastric tube or a plain red rubber catheter in the upper esophageal pouch. Evaluation of these neonates by a sagittal CT scan of the chest using the technique as described by Tam et al. was contemplated [10]. However, three neonates were too moribund to be moved to the CT scanning room. In another two other neonates, the study remained incomplete because of sudden deterioration in their condition on the CT table following aspiration of refluxed gastric secretions through a wide fistula. In 10 remaining consecutive neonates, chest CT scans were possible. Incidentally, all of them had an atretic upper esophageal pouch and distal pouch fistula (type C, Gross classification). The CT scanner at our center is a 3rd-generation Wipro GE Sytec 2000 with a gantry diameter of 65 cm. The babies were restrained in the plane of gantry without any sedation. Normothermia was ensured by wrapping the neonate in a warm thermal blanket. A total of five sections—one midline and two sections each of 3 mm on either side of the midline—were taken. In cases where the sagittal CT scan films failed to depict the esophageal anatomy clearly, axial cuts focusing on the area of interest were also taken; for example, in two patients with dextrocardia. Axial cuts were also taken in two patients in whom the upper pouch was thought to be short. The CT scan films were carefully evaluated for the type of esophageal atresia, vertebral level of the lower extent of the upper pouch, origin and level of the fistula, interpouch distance, and status of the heart and lungs. These details were then compared with the oper-

ative findings. A right posterolateral thoracotomy was performed in all the neonates except two cases with dextrocardia who had left bronchoesophageal fistula. Intraoperatively, the interpouch distance was measured following initial dissection to define the pouches, but before opening them. A silk thread between two artery forceps was used to define the gap between the lowermost limit of the upper pouch to the highest point of the lower esophageal pouch. This distance, in centimeters, was measured with a ruler. The distance between the pouches was measured again after division of the fistula. Single-layer full-thickness esophageal end-to-end anastomosis using 5-0 Vicryl sutures was done in all cases except for the one with type 4 (Benjamin and Ingelis) laryngotracheoesophageal cleft (LTEC). No mobilization of the lower pouch was done. In the neonate with LTEC, only division and ligation of the fistula, along with a feeding gastrostomy, was done initially, because of the patient’s poor general status. In two neonates with interpouch gap longer than 1.5 cm, extensive mobilization of the upper esophageal pouch up to the cervical region was done, and circular myotomy was also performed in one of these neonates. Postoperative ventilatory support was needed for both of these neonates and also for the one with LTEC.

Results The results of the study were computed from 10 consecutive neonates in whom CT scanning could be done before surgery. All of them had type C esophageal atresia (Gross classification); that is, an atretic upper esophageal pouch with distal pouch fistula. Dextrocardia was present in two neonates, and one of them also had a complex cardiac anomaly that could be identified postoperatively only. The other four neonates additionally had anorectal malformations. One neonate had an absent radius on the right side. Clinically occult vertebral anomalies were found in three neonates. By direct sagittal chest CT scan, two out of 10 were found to have left bronchoesophageal fistula (an unusually high incidence), and one had type 4 LTEC. The lower extent of the upper pouches reached the level of T2 vertebra in two patients, T3 in three, and T4 in five. At presentation, a moderate degree of pneumonia was present in three neonates. The neonate with LTEC had fulminant chest infection. Yet another patient with dextrocardia and long-gap atresia developed fulminant chest infection in the postoperative period. Sagittal CT scanning of the chest could depict the exact midline anatomy of the esophagus and trachea in relation to the origin of the fistula in most cases (Fig. 1), as was confirmed operatively. However, we realized that exact central positioning was vital for correct interpretation of the findings. Unfortunately, LTEC in a neonate with EA was missed because of improper positioning of the patient and the angulation of the cuts, which gave a false impression of the existence of inter-

759

responding to the intraoperative finding of overlapping esophageal pouches) to a maximum of 1.82 cm. Perioperatively, these findings corroborated in all the cases within range of ±2 mm. The primary anastomosis was relatively easily accomplished in all the neonates except in two in whom interpouch distance exceeded 1.5 cm on preoperative CT films. The effective gap to be bridged in these cases was found to increase by approximately 5 mm after division of the lower pouch fistula.

Discussion

Fig. 1 Midsagittal section of thorax of a neonate showing distended proximal esophageal pouch (upper arrow) and distal esophagus (lower arrow) and distal pouch fistula communicating with trachea anteriorly (curved arrow).

posing tissue between the esophagus and trachea, thus masking the presence of the cleft. This was the 2nd case of our study. This failure not only made us more cautious about exact sagittal positioning of the patient, but it also necessitated including axial cuts whenever the anatomy in sagittal films was not clear or whenever the upper pouch was detected to be unusually small. Subsequently, using this combined approach we were able to pick up left bronchoesophageal fistula in two neonates with dextrocardia, in whom the origin of the fistula in the midline CT film was unclear due to the presence of a cardiac shadow on the right side (Fig. 2a, b). The interpouch distance in preoperative CT films of the neonates with EA in our study varied from 1 mm (cor-

Fig. 2 a Sagittal section of thorax in a neonate with dextrocardia and esophageal atresia, showing widely separated esophageal pouches. The origin of the fistula is not clearly made out. b Axial section of chest of the same neonate at the level of the carina shows the origin of distal pouch fistula from the left main bronchus (arrow).

Clinically, a history of excessive salivation with inability to pass an esophageal tube beyond 8–10 cm from the angle of the mouth suggests a diagnosis of EA. Chest radiographs following the passage of a stiff radiopaque catheter in the esophagus still remain the investigation of choice at most centers for preoperative evaluation of neonates with EA [3, 5, 7]. A rough idea of the length of the upper esophageal pouch can be obtained by seeing the vertebral level of the catheter tip on the chest radiographs, with short upper esophageal pouches usually seen in cases with upper pouch fistula or LTEC [3, 5, 7]. This simple investigation also provides additional information regarding the existence of a distal fistula (by demonstration of abdominal bowel gas), the presence of other gastrointestinal and vertebral anomalies, and the neonate’s cardiac and pulmonary status. However, one cannot adjudge the level of origin of the fistula with the respiratory tree in most cases, though an air-filled column in the lateral view may hint about the same in some patients. The origin of the fistula can vary widely, from 0.5–1 cm from the carina (the carina lies opposite the upper part of the body of the T4 vertebra on chest radiographs) to the bronchus. The origin of the fistula from the left bronchus can mandate a left thoracotomy rather than the usual right-sided thoracotomy approach. The knowledge of origin of the fistula is helpful to a surgeon not only in deciding the side of thoracotomy but

760

also in anticipating the gap to be bridged. At the same time, the stretchability of the pouch by a stiff red rubber catheter can give a fallacious impression of a longer pouch length on chest radiographs, and an anastomosis initially thought to be ‘‘relatively easy’’ may turn out to be a difficult procedure requiring superior surgical expertise, a longer operative time, and postoperative ventilatory support. In addition, there is no provision to measure the gap between the upper and lower esophageal pouches preoperatively. Rare instances are cited in the literature in which a diagnosis of EA was masked when a thin tube put in the upper esophageal pouch passed down the trachea and TEF into the stomach [10]. Opinions differ widely on using a dye study of the upper pouch in patients with EA. Though a dye study may be conducted to delineate the upper pouch and a possible proximal pouch fistula, some researchers feel that the injection of air down an upper esophageal pouch and the consequent ‘‘airogram’’ may be a better alternative than contrast agents because of the higher risk of aspiration with the latter [5, 10]. This type of investigation also does not indicate the origin of the TEF or the interpouch distance. Preoperative endoscopy (esophagoscopy and bronchoscopy) prior to thoracotomy is routinely done in some centers to assess the length of the upper pouch and fistula [4, 8]. This technique is particularly useful for diagnosing an upper pouch fistula, H-type fistula, or the existence of a cleft. In experienced hands, the exact site of entry of the fistula can be correctly documented, but there is no objective assessment of the interpouch gap, and it requires an increase in the anesthesia time and its attendant morbidity and the need to change the baby’s position before the definitive surgical procedure. In addition, there is a danger of gastric distension and even rupture due to escape of anesthetic gases through a wide distal fistula [8]. The instruments and expertise for neonatal endoscopy are not widely available, especially in developing countries such as ours. Direct sagittal CT scan of the chest, on the other hand, is a simple, noninvasive, and quick means of preoperatively getting the vital information regarding the TEF and level of pouches in neonates with EA [10]. The most important advantage of this modality is that the esophagus and trachea are seen in their natural (unstretched) positions, and the exact interpouch gap can be measured on CT films by an inbuilt scale [10]. Tam et al. found preoperative interpouch distances to corroborate with operative findings in both their neonates under study. Our findings compared favorably with their observations. However, we realized that exact positioning is vital for correct information or else it can lead to a false impression of interposing tissues, as happened in one of our patients with LTEC (type 4). Guided by this failure, we started taking axial cuts in the area of interest whenever either the origin of the fistula or relevant anatomy was not well defined on a sagittal CT film [12]. On axial CT films in two patients, the fistula was found to be originating from the left main

bronchus. The preoperative information derived from sagittal and axial CT studies was highly useful because much time was saved in identifying and reaching the fistula. In addition, we could keep postoperative ventilatory support ready in view of long-gap in both these cases. In our study none of the patients had an upper pouch fistula or pure EA. Because distension of the esophageal pouches with air is essential to pick up the requisite information on chest CT films, this modality is likely to yield poor results in variants of EA where the distal esophageal pouch remains collapsed. However, once air is introduced into the gastrointestinal tract after a gastrostomy, this technique may be better than conventional techniques of monitoring interpouch distance before delayed primary repair [2, 5]. Maintaining normothermia and preventing aspiration of secretions remain other challenges during the CT scan. The advances in CT technology with quicker and better imaging should overcome these shortcomings. Our preliminary study shows that chest CT scan is a quick means of providing valuable preoperative information in neonates with EA. In our study, this preoperative information helped us (1) decide whether the repair should be done by an experienced surgeon or by a trainee with/without supervision, (2) anticipate the need for postoperative ventilatory assistance (based on the gap between the pouches and the lung status), (3) decrease the anesthesia and surgery time and its attendant morbidity, and (4) prognosticate a case, especially when the information on the interpouch gap was combined with the neonate’s general state. In view of the meager ventilator facilities available at many centers in our country, this preoperative decision was crucial for postoperative management. Although interpouch gap is considered an important predictor of long-term outcome in neonates with EA and TEF, there is no general consensus as to how this distance has to be assessed preoperatively—natural or stretched position—and between which points the distance should be measured—from the origin of the fistula or from the apex of the curvature of the lower esophageal pouch. In this study we measured the shortest straight distance between the air columns of both pouches preoperatively, as seen on CT films. Perioperatively, the shortest distance between the lowermost end of unmobilized upper pouch and the highest point of lower pouch curve (the shortest interpouch distance) was found to increase by as much as 5 mm following fistula division. Of late, there is greater enthusiasm to mobilize the lower pouch down to the hiatus in order to accomplish primary esophageal anastomosis, but long-term studies regarding the result of this approach are yet awaited. While plain chest radiographs are still the mainstay of investigating patients with EA and TEF, we feel that sagittal chest CT scan should be employed for the reasons stated above. Every center, however, has to work out the cost-effectiveness of this approach. Further

761

research needs to be done to establish and define the utility of this modality for other variants of EA also. Acknowledgements The authors are indebted to CSIR for financial support for the preparation of the manuscript.

References 1. Brown AK, Tam PK (1996) Measurement of gap length in esophageal atresia: a simple predictor of outcome. J Am Coll Surg 182:41–45 2. Caffarena PE, Mattioli G, Bisio G, et al. (1993) Long gap oesophageal atresia: a combined endoscopic and radiologic evaluation. Eur J Pediatr Surg 4:67–69 3. Cudmore RE (1990) Oesophageal atresia and tracheo-oesophageal fistula. In: Lister J, Irving IM (eds) Neonatal surgery. Butterworth, London, pp 231–258 4. Filston HC, Rankin JS, Grimm JK (1984) Esophageal atresia—prognostic factors and contribution of preoperative telescopic endoscopy. Ann Surg 199:532–536 5. Guiney EJ (1996) Oesophageal atresia and tracheo-oesophageal fistula. In Puri P (ed) Newborn surgery. Butterworth-Heinemann, Oxford, pp 227–236

6. Hands LJ, Dudley NE (1986) A comparison between gaplength and Waterston classification as guides to mortality and morbidity after surgery for esophageal atresia. J Pediatr Surg 21:404–406 7. Harmon CM, Coran AG (1998) Congenital anomalies of the esophagus. In O’Neill JA, Rowe MI, Grosfeld JL, Fonkalsrud EW, Coran AG (eds) Pediatric surgery. Mosby, Baltimore, pp 941–968 8. Holschneider AM, Gharib M, Ebel KL (1990) Morphological findings, classification and diagnostic procedures in esophageal atresia. In: Willital GH, Fekete CN, Myers N (eds) Management of esophageal atresia. Urban & Schwarzenberg, Munich, pp 20–25 9. Spitz L, Kiely EM, Morecroft JA, Drake DP (1994) Oesophageal atresia: at risk groups for the 1990s. J Pediatr Surg 29:723–725 10. Tam PKH, Chan FL, Saing H (1987) Diagnosis and evaluation of esophageal atresia by direct sagittal CT. Pediatr Radiol 17:68–70 11. Tam PKH, Chan FL, Yeung CK, Saing H (1987) H type tracheo-esophageal fistula. Pediatr Radiol 17:509 12. Wilkinson AG, Mackenzie S, Hendry GMA (1990) Complete laryngotracheo-oesophageal cleft: CT diagnosis and associated abnormalities. Clin Radiol 41:437–438