Evaluation of the blood flow in reconstructed gastric tube and its relation to anastomosis leakage

Anastomotic leakage is one of three major complications in esophageal cancer surgery [1]; however, only a few studies have evaluated the blood flow in reconstructed organs to reduce its frequency. The blood flow is evaluated by measuring the arterial blood flow in the gastric tube (GT) using visualization by indocyanine green (ICG) fluorescence during surgery [2]. ICG fluorescence can be evaluated qualitatively, but not quantitatively, making it difficult to evaluate venous return.

Another way to evaluate reconstructed GT blood flow is through regional tissue oxygen saturation (rSO2), which is commonly used to evaluate brain blood flow. rSO2 indicates oxidized hemoglobin/total hemoglobin, showing the balance between oxygen supply and consumption [3, 4]. Therefore, rSO2 can be used to quantitatively measure the blood flow which has not been possible before and may show its relationship with anastomotic leakage. Additionally, the measurement of the rSO2 values is simple, easy, and noninvasive, with lower rSO2 values and a higher decreasing rate of rSO2 may be associated with anastomotic leakage.

This study aimed to evaluate the blood flow before and after the formation of a reconstructed GT using regional saturation of oxygen measured by INVOS™ and determine its relationship with anastomotic leakage.

Patient characteristics are summarized in Table 1 based on the presence or absence of anastomotic leakage. Anastomotic leakage occurred in the middle thoracic esophagus in all cases (5 out of the 50 patients (10%), wherein four were men and one was women). The mean age of the patients was 68 years (51–82 years). We also identified docetaxel + cisplatin + 5-fluorouracil therapy (leak: n = 3, no leak: n = 27), cisplatin + 5-fluorouracil therapy (leak: n = 1, no leak: n = 7), others (leak: n = 0, no leak: n = 4), or absence (leak: n = 1, no leak: n = 7) of neoadjuvant chemotherapy and retrosternal (n = 4) or posterior mediastinum reconstruction route (n = 1).

The rSO2 measurements before GT formation at points X, Y, and Z were defined as X1, Y1, and Z1, respectively, while those after GT formation at points X, Y, and Z were defined as X2, Y2, and Z2, respectively.

The rSO2 measurements at each of the three points are shown in Fig. 2a–c. rSO2 at X2 and Z2 was significantly lower in the presence of leakage than in the absence (X: p = 0.03, Z: p = 0.02). However, there was no significant difference in rSO2 at X1 (p = 0.76), Y1 (p = 0.63), Z1 (p = 0.66), and Y2 (p = 0.62) between the presence and absence of leakage (Fig. 3). At point Z, the decreasing rate of rSO2 was significantly higher in the presence than in the absence of leakage (p = 0.01), but at points X (p = 0.052) and Y (p = 0.83), there was no significant difference in the decreasing rate of rSO2 between the presence and absence of leakage (Fig. 4). Multivariate analysis using logistic regression analysis showed a significant difference only in the decreasing rate of rSO2 in point Z (Table 2). The cutoff values for the decreasing rate of rSO2 and the incidence of anastomosis leakage at point Z would be obtained using ROC curves, the cutoff decreasing rate of the rSO2 was 29.41% (sensitivity: 80%, specificity: 98%, AUC: 0.94; Fig. 5).

GT reconstruction is the most common type of esophagectomy, which can have serious complications, such as anastomotic leakage, delayed oral intake, and prolonged hospitalization [1]. A narrow GT is usually used because of its association with good elevation and blood flow [5], with the latter being a very important factor in complete anastomosis [6‐8]. Thus, various studies have reported the methods of blood flow evaluation for GT formation.

For instance, the Doppler blood flow meter technique [9‐11] can evaluate GT blood flow by measuring the arterial blood flow in the gastric wall. The evaluation of blood flow using the ICG fluorescence imaging method has been generally reported in surgery [12‐15]; accordingly, it has also been used to evaluate GT blood flow. The ICG method has shown good results in qualitative evaluation, and some studies have shown an impact on the incidence of anastomotic leakage with and without qualitative evaluation [13, 14]. However, both venous and arterial blood flows are involved in anastomotic leakage. Fujioks et al. reported that venous pressure is weaker than arterial pressure, and the gastric tube passes through a narrow pathway, so the distal end of the gastric tube is often congested due to impaired venous return. If congestion continues, irreversible tissue injury can occur due to the reduced tissue tension of oxygen [15]. Thus, evaluation of arterial blood flow and venous return is synonymous with quantitative evaluation by ICG fluorescence imaging, but this has not been shown to be associated with anastomotic leakage [16]. In addition, Yukaya et al. attempted a quantitative evaluation using ICG in the context of fluorescence and extinction flow between two points in gastric tube. The patients were classified into normal type, inflow delayed type, and outflow delayed type. However, there was no significant difference in anastomosis leakage in each group. This may indicate that ICG is good at visually assessing inflows but not accurately assessing outflows. [17] Furthermore, the narrow GT is mostly supplied by the right gastroepiploic artery and vein, with 20% of the tip blood flow in the GT being provided by the capillary vessel of the gastric wall. Therefore, it is necessary to quantitatively evaluate the arterial blood flow and venous return including capillary vessels. To this end, INVOS™ using NIRS can be used to measure rSO2 because it indicates microscopic arterial and venous oxygen saturation, and the state of regional perfusion is represented by capturing changes in the regional oxygen supply balance [18, 19]. Recent reports state that NIRS can measure regional blood flow easily, noninvasively, and repeatedly [20‐22]. In the field of plastic surgery, Steele reported the usefulness of NIRS for the evaluation of ischemia in the construction of free flaps [20]. Additionally, the usefulness of NIRS has been reported in flap reconstruction and limb ischemia in vascular surgery. The cutoff score for tissue ischemia was 45% in these reports [21].

Yamaguchi et al. used the new noninvasive blood flow evaluation device “toccare” (Astem Co., Kanagawa, Japan) to evaluate blood flow in the GT [4] and found no significant differences in rSO2 values before and after GT formation; however, the relationship between anastomotic leakage and rSO2 could not be elucidated [4, 22].

In this prospective observational study, we used a similar device to evaluate the venous return and arterial blood flow in the reconstructed GT. The measurement of the rSO2 values was simple, easy, and noninvasive, and we used a narrow GT for good elevation. However, tip blood flow in a narrow GT might be poor. Blood flow and tension are major factors in anastomotic leakage. The rSO2 values at each of the three points were significantly lower before GT formation than after, which indicates that GT formation may reduce blood flow in the gastric wall.

The rSO2 values at the antrum (point X) and anastomotic point (point Z) were significantly lower with anastomotic leakage than without. No significant differences were detected in the rSO2 values at point Y with and without anastomotic leakage. We believe these results were influenced by the lack of reference or normal value of rSO2 for each organ and tissue. In some cases, a high rSO2 might have decreased after GT formation, whereas a low rSO2 might have remained low after GT formation. The decreasing rate of rSO2 was significantly different only at point Z. Given the results of the multivariate analysis, the decreasing rate of rSO2 rather than the actual rSO2 values is more likely to represent congestion or ischemia. This rate at an anastomotic point was significantly higher only before GT formation than after. No study has ever described the decreasing rate of rSO2, and this is a new finding. The decreasing rate of rSO2 may be more important than the rSO2 value as a predictor of anastomotic leakage, and it is recommended that anastomosis be performed at sites with a lower decreasing rate. Anastomosis may reduce anastomosis leakage by performing the sites with the decreasing rate of rSO2 less than 29%. It is also suggested that the addition of superdrainage be considered if the gastric tube is shortened.

This study has limitations. First, the sample size was too small to determine whether the rSO2 values were related to anastomotic leakage. However, the decreasing rate of rSO2 before and after GT formation showed sufficient potential to predict anastomotic leakage, which has not been explored before. Second, the rSO2 values were unstable. It is clear that GT formation decreases GT blood flow. However, we observed cases in which rSO2 decreased at point X, which was thought to be unaffected by reduced blood flow, and rSO2 increased after GT formation. The decrease in rSO2 at point X was thought to be due to the increase in reduced hemoglobin caused by venous stasis due to the decrease in venous return caused by GT formation. However, the increase in rSO2 after GT formation was thought to be due to measurement instability. The different conditions at the time of measurement in each case and the fact that the measuring probe is equipped with a probe cover made the values unstable. To solve these problems, efforts were made to obtain value stability by taking the measurement time. The decrease in rSO2 at Point X after gastroduodenal tube formation is thought to be due to an increase in reduced hemoglobin level due to venous stasis.

In the future, we will analyze more cases to investigate the relationship between rSO2 and anastomotic leakage. The need to increase the credibility of statistical considerations was discussed.

We evaluated the rSO2 of GT to determine the relationship between GT blood flow and anastomotic leakage. rSO2 may be useful for measuring blood flow and the decreasing rate of rSO2 may be a predictor of anastomotic leakage. Maintaining consistent conditions for rSO2 measurements can be challenging due to the sensitivity of rSO2 to changes in circulating blood flow and oxygen administration. In order to establish more detailed and appropriate rSO2 criteria, an additional and larger study is needed.