The present study demonstrated that the cut-off value of the gap length at which primary anastomosis can be performed without complications was < 2.0 cm, which had a specificity > 90%. Thus, a long gap may be defined as being 2.0 cm or more in size. Most of the patients had a good, long-term prognosis and improved without oral feeding difficulties over the long term regardless of gap length even when anastomotic leakage occurred. Patients with long-gap will probably require therapeutic intervention for postoperative complications and an extended postoperative management period. Nevertheless, the primary anastomosis was performed successfully in all the EA patients in the present cohort with a gap ≤ 3.0 cm. Therefore, a primary anastomosis should be performed for EA. However, patients with a gap length > 2.0 cm should be carefully managed postoperatively, and informed consent should be obtained before treatment because complications are unavoidable. Additional procedures, such as drain insertion in case of leakage, fundoplication or anastomotic dilatation, might be required.
Several studies have specifically defined a long gap based on personal experience [21] or previous research [17, 19], but as of yet there is no consensus on a standard length for defining LGEA because the definition of a long gap has varied across studies. A long gap is often defined functionally as having a length precluding a primary anastomosis or leading to a failed anastomosis [3]. However, this definition is imprecise because the performance of a primary anastomosis depends heavily on the skill of the surgeon and other variables [4]. Specifically, Spitz et al. [4] suggested that if a gap is greater than six VB, the chance of saving the esophagus is remote while Friedmacher et al. [8] stated that esophageal replacement was necessary only in a few patients who either had no distal esophageal segment or only a nub of the distal esophageal segment. Additionally, Bagolan et al. [21] reported that nearly all the patients in their study were able to be anastomosed successfully using any of several maneuvers, such as extensive proximal or distal mobilization, dissection of the upper esophageal pouch via a cervical incision, internal traction on both segments or the addition of an upper esophageal flap. In the present study, early primary anastomosis was able to be performed for EA with a gap length of up to 3.0 cm.
Some studies not providing a specific gap length in their definition of LGEA have defined LGEA by the presence of a distal fistula at the carina [6, 22] or the absence of a TEF [23] other than by the failure of anastomosis [1, 24–26].
On the other hand, in studies defining gap length in centimeters, lengths including > 2 cm [9, 10, 27–29], > 2.1 cm [30], ≥ 2.5 cm [15, 19], > 3 cm [31], and ≥ 3.5 cm [32], and ≥ 5 cm [33] were use to define a long gap. Studies using VB defined a long gap as VB ≥ 2 [4, 34], ≥ 3 [21], ≥ 4 [35] or ≥ 5 [36]. Several studies have employed two definitions, namely, VB > 3 cm or > 2 [17, 37]; ≥3 cm or ≥ 5 [16]; or ≥ 3 and/or the inability to perform the primary anastomosis in the first operation [13]. Other articles have included a subset of LGEA dubbed ultra-LGEA, which has a gap length ≥/>3.5 cm [19, 38] or > 6 cm [39]. Al-Shanafey et al. [1] reported that 40% of pediatric surgeons they surveyed defined LGEA as having 3–4 VB, and 24% thought it had no TEF. Baird et al. [2] reported that an audience response system poll revealed that 63% of physicians favored using VB to describe gap length while 10% simply ‘know it when they see it’.
In EA, the gap length between the proximal and distal segments is a measure of the severity of the anatomical defect and is directly related to the degree of technical difficulty of surgery [19, 31]. The longer the gap, the greater the tension if primary end-to-end esophageal anastomosis is carried out [19]; therefore, a long gap is associated with an increased incidence of anastomotic complications, such as leakage, strictures, and GERD [6, 10, 31, 40]. LGEA is challenging for pediatric surgeons. Because the survival of infants with EA has improved markedly, the relationship between gap length and both short and long-term outcomes has begun to attract attention [22, 27, 29]. It has been known for several decades that gap length largely determines the complication rate [27, 28, 38]. When a gap is 2 to 3 cm or more, the complication rate is significantly higher [28, 31, 41]. On the other hand, there are surprisingly few studies reporting a detailed comparison of outcomes based on gap length [9, 10, 27, 29, 31]. Sillen et al. [29] found that LGEA (gap length > 2 cm) was associated with more anastomotic complications than no or moderate gap-EA. Hands et al. [27] reported that postoperative complications, such as anastomotic leakage, wound sepsis, septicemia, pneumothorax, and pneumonia, were significantly more common in patients with LGEA (gap length > 2 cm). In a more recent study dividing patients into three or four groups according to the gap lengths documented during thoracotomy, Brown et al. [31] demonstrated that the incidence of anastomotic complications was highest in the long gap group (> 3 cm) while it was moderate in the intermediate gap group (> 1-≤3 cm). Upadhyaya et al. [10] divided EA patients with TEF into four groups (ultralong: >3.5 cm, long: 2.1–3.5 cm, intermediate: >1-≤2 cm, and short: ≤1 cm) and found a statistically significant difference in the incidence of esophageal leakage and mortality. Thakkar et al. [9] divided their patients into three groups (long: >2-≤5 cm, intermediate: >1-≤2 cm, and short: ≤1 cm) and found that gap length had no significant correlation with leakage or stricture; however, they observed a significant increase in the need for fundoplication. Orringer et al. [42] found GERD to be more common in patients who required a distal esophageal mobilization of 3 cm or more for the primary anastomosis.
Moreover, no consensus exists on how and when to measure gap length[1, 6, 21]. Intraoperative direct measurement [1, 9, 10, 31] and preoperative radiological measurement [11, 12, 17, 19, 20, 34, 37]) of a gap using VB or in centimeters have been reported. In EA without a distal fistula (type A or B), radiological measurement of gap length is generally performed after creating a gastrostomy. However, there are few reports describing a method of preoperatively assessing gap length in EA with a distal fistula in patients without a gastrostomy [21]. Bagalon et al. [21] reported measuring gaps with a combination of fluoroscopy and bronchoscopy. However, this method is not common. On the other hand, several studies have intraoperatively measured gap length during a thoracotomy [1, 19, 24]. The clinical significance of this method has not been substantiated in some reports because the studies lacked uniformity. Most studies have failed to indicate whether a gap was measured before or after dissection and if it was measured under tension [1]. The present study used direct, uniform, intraoperative measurement of gap length, in which the measurement deviations are negligible, as the most accurate and easiest method to perform [31].
Body weight reportedly affects esophageal strength [43]. Therefore, the impact of gap length on anastomosis may vary depending on the patient’s body weight. The present study found a significant difference in body weight at surgery. In low-birth-weight infants weighing less than 2.5 kg, a long gap tended to produce worse outcomes.
Our report has some limitations. The retrospective analysis might have limited the validity of the data. Intraoperative measurements of gap length were reviewed retrospectively, and repeated measurements were not taken as in a prospective study. Therefore, it was not possible to evaluate the reproducibility of this measurement method with statistical reliability. Different surgeons performed the surgery and managed the patients, thereby introducing the possibility of a measurement bias arising from individual variations in experience, etc.
Moreover, the patients who underwent delayed primary repair or esophageal replacement as the primary repair were excluded. Because patients with a gap length > 3 cm were classified as Gross type A or B and underwent delayed primary repair, primary repair of a gap length longer than 3.0 cm was not included. A prospective study of EA should be undertaken using uniform criteria to measure gap length and to compare results from different institutions.