Evaluation of preoperative risk factors and postoperative indicators for anastomotic leak of minimally invasive McKeown esophagectomy: a single-center retrospective analysis

Anastomotic leak is a common and serious complication among esophagectomy patients [1, 2]. With the development of both surgical technique and perioperative management, the incidence of AL has decreased over these years, especially in minimally invasive esophagectomy, occurring in 2–2.96% of patients undergoing Ivor-Lewis esophagectomy while 6.6–13.64% receiving McKeown esophagectomy [3, 4]. The mortality rate of AL decreases approximately to 1.7% at experienced centers [5]. However, anastomotic leak is still a tough problem for thoracic surgery as it results in prolonged hospitalization psychological distress and even an increase in cancer recurrence [6]. Besides, as gastric conduit is widely applied, most esophageal anastomotic leak, small and latent with varied clinical symptoms in early stages, could be likely to result in abscess cavity or even systematic infection when be found out [7]. Thus, even the anastomosis is close to the cervical incision in McKeown esophagectomy, the signs of leak are not always obvious enough for early recognition and management, resulting in serious infection in neck, mediastinum and even chest [8, 9]. We have collected preoperative variables to identify which are risk factors for anastomotic leak. Meanwhile, the postoperative indicators were recorded and analyzed to assess whether they could detect or predict anastomotic leak.

Minimally invasive esophagectomy becomes popularized rapidly, which means it limits tissue trauma without compromising oncologic resection [5, 13, 14]. However, as gastric conduit is routinely used for replacement of esophagus, gastric conduit failure appears to be more common than previous open surgery [10]. Besides, patients receiving minimally invasive McKeown esophagectomy are more vulnerable to gastric conduit failure than patients who received minimally invasive Ivor-Lewis esophagectomy [15, 16]. Once the anastomosis or the tip of gastric conduit undergoes ischemia, it may finally progress to anastomotic leak over a period. As crevice increases in size, the symptom develops from mild to serious. On early detection, this complication could be treated timely to avoid serious consequences.

Amylase is a digestive enzyme found in low concentration in blood (< 140 IU/L) but high in saliva (70,000 IU/L) [17]. The drainage amylase level has been used to diagnose anastomotic leak in laryngectomy, pancreatectomy, gastrectomy and Roux-en-Y procedures in past decades [18–20]. Thus, it is a reasonable hypothesis that amylase could also be used as an indicator for esophageal anastomotic leak by testing the concentration of the drainage fluid following esophagectomy [21, 22]. But in previous studies, the placement of drainage tube was high-demanding: locating next to the anastomosis with a specialized tube. Detecting anastomotic leak by drainage amylase for minimally invasive McKeown esophagectomy has never been recorded before. In the past we routinely put a drainage strip in the neck incision and removed it in 48 h, which was convenient for detecting anastomotic leak. However, anastomotic leak often occurs after POD 3 according to our experience.

The drainage tubes in our study were placed in the chest instead of next to the anastomosis by neck incision. We suppose the amylase of drainage fluid is also sensitive to anastomotic leak. Since the thoracic cavity pressure is negative, once the anastomotic leak occurs, saliva could be sucked into chest. By the movements of the breath, the drainage fluid would be contaminated.

Our study found that amylase cutoff of 55 IU/L on POD 6 was 100% sensitive and 86.96% specific in detecting anastomotic leak, and the corresponding AUC was the largest. However, there is no significant difference among the AUCs of POD 4, POD 5 and POD 6. According to principle of early diagnosis and early intervention, amylase cutoff value of 85 IU/L on POD 4 or 65 IU/L on POD 5 has higher diagnostic value. When these test results come along with symptoms such as fever, chest or incision pain and dyspnea, anastomotic leak should be taken into consideration. The reason for the difference of amylase between AL group and None AL group begins to show itself from POD 4 may be related with the course of digestive tract reconstruction. Doctor HU Xiang has divided this course into four stages: mechanical healing period(1-3d), pathologic inflammatory period(3-5d), tissue healing (fibrosis) period(5-7d) and maturation period(7d-later) [23]. In mechanical healing period, the joint of anastomosis depends on the stapler nails or sutures. When it comes to the inflammatory period, the joint begins to be strengthened by the tissue support force. Then in the healing (fibrosis) period granulation tissue starts proliferating and inflammatory cells are subsiding, causing mucous epithelium cells to grow and cover the anastomosis. After 7 days, the digestive tract reconstruction completes. POD 3–5 are just in the pathologic inflammatory period when the joint is weak and most anastomotic leaks occur (Fig. 5). Besides, gastric conduit necrosis is also an important factor causing anastomotic leak [5]. Doctor Darmarajah Veeramootoo [10] had researched on the relationship between gastric conduit failure and the postoperative C-reactive protein level. He also found that elevated CRP levels in the absence of any other clinical cause beyond POD 3 raises suspicion of incipient gastric conduit failure. Above all, POD 3–5 is a crucial period for doctors to recognize anastomotic leak and conduct appropriate management immediately before the impact of contamination develops. However, it’s supposed to be unsafe to received esophagram in this period, whereas drainage amylase is a good choice.

To confirm that the rise of drain amylase concentration is attributed to saliva from esophagus instead of pleural effusion, we collect the drain fluid from patients receiving pulmonary surgery on POD 1 and POD 2. We found that the drain amylase concentration in AL group is not only significantly higher than that in None AL group but also higher than the pleural fluid from patients of pulmonary surgery, while the levels of amylase concentration in both pulmonary surgery group and None AL group are similar. This result verifies our supposition that the distinct rise of drain amylase concentration in AL group is due to saliva, a fluid full of amylase [24], which mixes into pleural fluid.

Besides, the drop of drain amylase concentration on POD 6 in AL group draws our attention. We think that the significant decline of amylase level may be attributed to the clinical intervention. Many patients may have clinical symptoms such as fever, incision pain or elevation in white blood cell on POD 5 or POD 6 due to anastomotic leak. Therefore, treatments like incision flush and drainage may be conducted and the level of drainage amylase concentration consequently declines. These factors should be considered when interpreting our data. So, the relevance between the incidence rate of anastomotic leak and drainage amylase concentration may decline on POD 6.

Prealbumin is a protein synthesized in the liver, metabolized and excreted by the kidneys. As it has a short half-life, its serum level change rapidly in response to nutritional status. Once the patients got hypoproteinemia, the gastric conduit and bowel mucosa may appear edema consequently, which would worsen the absorption of nutrition and finally lead to anastomotic leak. Our results revealed that patients in AL tend to show lower prealbumin since POD 4 to POD 6. It’s understandable that surgery may significantly influence nutrient metabolism. In Ying-Jian Wang et al’s study [25], they recorded the preoperative and postoperative levels of albumin and prealbumin of patients who underwent MIE(minimally invasive McKeown esophagectomy). They found that the decrease of prealbumin after MIE was associated with the incidence of cervical AL. Interestingly, albumin was not observed to have such relationship with AL. Our study found that patients may have high risk of AL if their serum prealbumin concentration was below 128 g/L on postoperative day 5. However, low prealbumin concentration can’t reflect AL promptly, it may not be sensitive enough as an indicator of AL.

In our study, we also screened the risk factors of esophageal anastomotic leak among the patients’ characteristics, tumor features and therapeutic methods. Although previous studies have reported several hazards for AL [3, 26–28], we found no statistically significant risk factors in our study. As a retrospective study in single center with small quantity of patients, our study may have potential bias. It may reduce the reliability of cutoff value. Therefore, further prospective studies are needed to be done.

In summary, our study reveals that drainage amylase concentration on postoperative days may help to discover anastomotic leak in early stages after minimally invasive McKeown esophagectomy. Patients with amylase level above 85 IU/L on POD 4 or 65 IU on POD 5 should be considered suffering anastomotic leak especially when they have symptoms such as fever, incision or thoracic pain. Besides, prealbumin concentration below 128 g/L on POD 5 might be potential risk factor for anastomotic leak.