There is no correlation between a delayed gastric conduit emptying and the occurrence of an anastomotic leakage after Ivor-Lewis esophagectomy

816 patients were included in the study. A total of 226 (27.7%) patients were diagnosed with postoperative DGCE. Patients with postoperative DGCE had a significantly higher rate of postoperative pneumonia (13.7% vs. 8.5%, p = 0.025) and a prolonged hospital stay (median days 17 vs. 14, p < 0.001). The risk of postoperative DGCE was significantly higher in female compared to male patients (34.4% vs. 26.2% vs., p = 0.040). Table 1 depicts patient characteristics of patients with and without DGCE.

47 patients (5.8%) were diagnosed with postoperative anastomotic leakage ECCG type II. 769 patients out of this selected cohort (94.2%) had no anastomotic leakage. The two patient groups did not differ regarding baseline clinical characteristics like sex, age, ASA and histology. Furthermore, there was no difference in patients that underwent neoadjuvant treatment in both groups. Baseline characteristics for patients with and without anastomotic leakage are depicted in Table 2.

To test our hypothesis that patients with DGCE are associated with a higher risk for the occurrence of AL than patients without DGCE after esophagectomy, we investigated both collectives for AL. AL occurred in 5.8% of patients with DGCE and had an identical rate of 5.8% in patients without DGCE, respectively. In consequence, DGCE was not associated with AL Type II in our collective.

We analysed other potential factors which might have a possible impact on the occurrence of anastomotic leakage. First, the American Society of Anaesthesiology (ASA) Score, classifying patients in terms of their perioperative risk with respect to their preoperative condition [30]. We subdivided our patients in one group with patients classified as ASA I or II and compared them to a group of patients classified as ASA III or IV, assessed during preoperative work up by the visiting anaesthesiologist. ASA III and IV patients had a higher risk to suffer from AL (7.6%) than patients with ASA I and II (4.4%, p-value 0.05). We were furthermore able to show, that patients with a squamous cell carcinoma (SCC) of the esophagus have a significantly higher risk of AL compared to patients with adenocarcinoma (9.8% vs. 4.7%, p = 0.01). R0 resection rates for AC and SCC were comparably high (98% vs. 97%) and therefore not causally associated to the occurrence of DGCE after surgery.

In line with this, as depicted in Table 3, the logistic regression analysis revealed that in our collective the risk to develop AI type II is reduced by around 50% in patients with AC (OR 0.45, 95%-CI 0.24–0.83, p-value = 0.011) compared to SCC as well as in patients with ASA score I or II compared to ASA score III or IV (0.55, 95%-CI 0.30–1.00, p-value 0.05).

Our data clearly demonstrate that delayed gastric conduit emptying after Ivor-Lewis esophagectomy is a common functional disorder. DGCE can likely result in high morbidity with increased rates of postoperative pneumonia followed by a prolonged hospital stay. In our collective there was no association between a DGCE and an increased risk for anastomotic leakage. Though this is a retrospective single-center analysis our work has several strengths: (i) we analysed over 800 cases within 5 years from a prospectively maintained database (ii) the operative technique was identical in every case and was highly standardized (iii) our collective is highly representative as associations between DGCE and pneumonia e.g., could be retraced. To our knowledge, this is the largest cohort of patients examined for DGCE after a standardized treatment. With a DGCE rate of 27.7% in our collective the results are well comparable to international data [14]. Since ECCG Type I AL is a rare finding without a clinical or therapeutic relevance and thus without statistical impact they were no subject to this analysis. AL type III was usually diagnosed on the second postoperative day, before oral intake was started and therefore no diagnostic work up regarding DGCE was performed, these cases missed any significant statistical contribution to this study and were, therefore, excluded. Furthermore, patients with AL type III had clear signs of a not sufficient blood supply of the gastric conduit, which explained the cause for anastomotic leakage itself and, therefore, DGCE was not relevant at all.

One simple explanation for differences seen between the association of DGCE and AL compared to previously published results from others is that several publications include patients operated with different surgical techniques (including McKeown procedure, cervical anastomosis) that might have biased the results [25, 31]. Another potential bias might be, that DGCE usually is treated endoscopically, increasing the possibility to detect small, clinically inapparent anastomotic leakages during the endoscopic procedure [29]. A recent report from Hadzijusufovic et al. evaluating the usefulness of a preoperative endoscopic balloon dilatation as a strategy to prevent postoperative DGCE indirectly supports our findings. The authors did not find a significantly decreased rate of anastomotic leakage or early pulmonary complications in their intervention group (preoperatively dilatated pylorus) [32]. Indeed, the latter work demonstrated a significantly lowered rate of DGCE in patients that underwent preoperative dilatation of the pylorus during staging/restaging endoscopy compared to patients who did not receive an endoscopic pneumatic dilatation of the pylorus prior to Ivor-Lewis esophagectomy. Still, a preoperative dilatation usually is performed with a 20 mm balloon, whereas a postoperative dilatation can be performed with a 30 mm balloon, which was initially designed for the treatment of achalasia patients. Thus, a re-dilatation was more often necessary in patients after a 20 mm balloon procedure. Randomized clinical trials that evaluate the interventional option of preoperative dilatation or postoperative medical therapy are currently lacking in the literature and should clearly be considered in the future. At least a partial recovery of the gastric peristalsis has been reported earlier, which might even improve by medical therapy [33‐35]. This is in line with reports of a regaining of gastric acidity in the denervated stomach [36]. Surgical perioperative pyloric drainage procedures like pyloromyotomy or pyloroplasty seem to significantly lower the risk of DGCE but without any impact on other morbidities such as AL and/or pulmonary complications, as reported in the meta-analysis by Urschel et al. [37]. In addition as Palmes et al. described surgical pyloric procedures are reported to contribute to a higher rate of bile reflux and esophagitis without having a significant benefit in the early postoperative phase, even though there seems to be a trend towards less early pulmonary complications and a lower rate of anastomotic leakage [38]. This trend is in line with the data published by Hadzijusufovic, as described earlier [32]. This might be the case, since the surgical drainage procedures are more invasive and usually a definitive therapeutic option compared to the endoscopic balloon dilatation, which sometimes even has to be repeated for a successful therapy. In summary, there is no high-quality evidence which supports pre- or perioperative pyloric drainage procedures to significantly contribute to a reduction of postoperative morbidity, even though the rate of DGCE can be reduced. Obviously, the interrelation of DGCE and postoperative morbidity is not finally understood and requires further investigation.

Another issue that has to be addressed is that there remains a small group of patients with long-term complaints associated to DGCE being refractory to repeated pyloric dilatation. This is a very demanding situation and depends on individual clinical properties of each patient. In these patients a pyloric spasm does not seem to be the reason for DGCE, more often it is a hiatal narrowing or a kinking of the gastric conduit, which results in a kind of gastric outlet obstruction. Treatment strategies in our own collective range from conservative prokinetic drug treatment through endoscopic pyloric stenting to surgical revision. More standardization of this clinical problem is further needed to define treatment recommendations. There is a lack of literature for this rare but regularly occurring problem. At least, a consensus about the definition of DGCE found broad acceptance and therefore benefits to a better comparability of international studies [28].

Overall, our work underscores that DGCE has to be considered as a frequent complication after Ivor-Lewis esophagectomy and a reconstruction with a gastric conduit pull-up. It significantly contributes to postoperative morbidity, is primarily associated to an increased rate of pulmonary complications and a prolonged hospitalization. According to our findings DGCE can be seen as its own entity of postoperative morbidity. In the absence of other clinical or endoscopic signs DGCE should not be a direct warning sign for an accompanied anastomotic leakage making it an important finding for treatment pathways of affected patients. Clinical and/or radiologic suspicion of DGCE in the absence of signs of infections does not require urgent endoscopic intervention which might have its own risks. Usually, a drainage with a nasogastric tube resolves most clinical symptoms and gains time for a planned endoscopic therapeutic procedure under controlled and safe circumstances.