This study examined the effect of our novel concept of indwelling an IAD suction tube to prevent POPF during PD reconstruction. This method is expected to reduce undesired pancreatic juice from the branched pancreatic duct and duct-to-mucosa anastomosis and compensate for the shortcomings of Blumgart-type pancreatojejunostomy. This method successfully collected notably amylase-rich fluids in the inter-anastomosis space. However, we could not identify meaningful improvements in terms of the incidence of clinically relevant PF and POHS duration.
Our IAD tube method showed, for the first time, the existence of a substantial volume (maximum of 14.0 mL/day [median] in total 86 patients from the B+IAD group and 16.5 mL/day in patients with a soft pancreas, Table 3) of undesirable fluid in the interspace between the jejunal wall and pancreatic parenchyma that contained markedly high amylase levels (median value: 90,900 IU/L in patients with a soft pancreas). This undesired fluid could be considered the origin of PF. However, it is difficult to ascertain whether fluid effluxed from the branch duct was a minor leakage from the duct-to-mucosa anastomosis. From this result, we suggest that the soft pancreas produced a high amount of amylase-rich pancreatic juice and the operator’s subjective palpation during the procedure regarding whether the pancreas was soft or hard seems correct.
Although the IAD tube successfully collected fluid with notably high amylase levels at the inter-anastomosis space in Blumgart-type pancreatojejunostomy, we could not identify any meaningful improvements in clinically relevant PF. PF grades B and C occurred in 28.1% and 5.3% of patients, respectively, in the B+IAD group and in 31.9% and 3.4% of patients, respectively, in the B group. Our initial report on 17 patients using this novel IAD tube method revealed a significant reduction in the incidence rate of PF grade B or C (the incidence rate was 5.9%; 1 patient out of 17 patients) 15. The IAD tube functions effectively when the pancreatic cut margin is wide and the MPD is located on the dorsal side (Fig. 3a). Conversely, when the surface area of the pancreatic stump is small or the MPD is located ventrally (Fig. 3b, c), the insertion of IAD tube is troublesome and the result is undesirable.
The reason why the IAD tube method did not improve after the first report may be due to three reasons. The first reason is the technical difficulty in the placement of the IAD tube at the inter-anastomosis space between the jejunal wall and the pancreatic parenchyma, especially in cases including a thin pancreas stump. For the IAD tube to exert a pronounced effect in order to protect leakage from the branch duct, the tube needs to be placed in a water-tight-seal manner. This water-tight seal was made using 2 or 3 Blumgart suture knots at the jejunal surface, while embedding the IAD tube above the duct-to-mucosa anastomosis 15 (Fig. 1b, c). Moreover, this is occasionally a highly challenging problem especially in the case of a small pancreatic cross-section. When the above side of pancreatic stump is thin and same as the width of IAD tube, the tube placement with a water-tight seal is theoretically impossible. In fact, the 17 patients reported in our first report 15 were operated by staff surgeons engaged in pancreatic surgery. This suture technique, creating a water-seal adhesion, is technically difficult for trainees. In fact, the operating time was longer in the B+IAD group than the B group in the soft pancreas cohort. This may reflect the reconstruction required more times to prudent suture techniques.
Second, the location of the MPD in the stump was important for effective collection. When the MPD was located on the ventral side of the stump, it was quite difficult to place the IAD tube above the duct-to-mucosa anastomosis similar to that in patients with thin parenchyma. Moreover, the fluid that leaks below the anastomosis might not be collected effectively by the IAD tube because the IAD tube was only placed above. This might also indicate that the IAD tube could not entirely collect the undesired fluid discharged from both the pancreatic stump and the incomplete duct-to-mucosa anastomosis.
The third reason was that the IAD tube placement may prevent tissue adhesion, which is important for the cure of PF. Though IAD successfully collected the amylase rich fluids, foreign substances may delay self-healing. The median maximum amount of IAD tube collection was 14.0 mL/day and more than 100 mL/day was recorded in some patients. This observation may suggest that the collections from the IAD tube included not only pancreatic juice but also enteric fluids that contained abundant enterobacteria. Though IAD tube placement may trigger bacterial infection around the pancreatojejunostomy, we did not experience any fatal infectious complications during the study period.
To overcome the technical difficulties associated with IAD tube placement, a simpler and more reasonable method is required to allow the pancreatic juice to drain into the jejunum. Recent studies have suggested that the healing of PF requires the amylase-rich fluid to drain spontaneously into the jejunal lumen 12. To achieve spontaneous drainage at an early stage after PD, the intestinal internal pressure needs to be kept low by the placement of an enteric drainage tube near the pancreatojejunostomy. High intestinal internal pressure may induce the outflow of enteric juice and bacterial infection, which activates proteases. Subsequently, the intentional drainage route in the jejunal wall may enable the fluid to spontaneously drain into the jejunum from the early postoperative period without the placement of an IAD tube. To accomplish this, two key factors, tight adhesion between the jejunal wall and pancreatic stump and low pressure inside the jejunum, are required.
In the present study, the incidence of clinically relevant PF (grade B or C) was 23% in the total cohort and 34% in the soft pancreas cohort. These values were relatively high compared to other studies on Blumgart-type anastomosis 6, 9. However, POHS did not change significantly compared to these studies 6, 8, 16. In fact, we achieved 17 days of POHS after PD without mortality. According to the ISGPF grading system 13, PF grade B requires a change in the postoperative management; thus, drains were either left in place for more than 3 weeks or replaced by endoscopic or percutaneous procedures. We generally performed an early assessment of the obstruction of drainage tubes on POD 3 or 4 and replacement of tubes under radioscopy if the drainage was not effective. Patients who underwent drain tube replacement or for whom the drainage site was changed under radioscopy were judged to have PF grade B. These facts may be responsible for the increased number of patients diagnosed with PF grade B in our study. Even if a patient experiences clinically relevant PF, the duration of POHS is important.
This study has some limitations. First, this is a retrospective study from a single institute, i.e., a teaching hospital. The sample size was limited, and each procedure was performed by several surgeons, including trainees. All procedures were guided by three staff surgeons; however, maintaining reproducibility was not easy due to the technical difficulty of IAD tube insertion. Second, the judgement of soft or hard pancreatic texture was made subjectively by the surgeons. Potential selection bias is difficult to eliminate, and a precise method to assess pancreas status is warranted. Third, pancreatic exocrine secretion depends on the size of remnant pancreas, acinar cell function, and number of detailed branch ducts in the stump. Patients with low exocrine potential may not require this IAD method. Therefore, further studies are necessary to clarify the relationship between remnant pancreatic function and the effect of IAD tube insertion.