Gastrostomy as a Preemptive Measure after Pancreatoduodenectomy against Delayed Gastric Emptying: A Small Case Series and a Review of the Literature

Delayed gastric emptying (DGE) is a common (20–30%) postoperative complication following pancreatoduodenectomy (PD) (Parmar et al., 2013). Various causes and preemptive measures have been suggested to decrease the occurrence of DGE. We added a simple step in the procedure of 26 consecutive pancreatic head resections, which seems to alleviate DGE and has never been highlighted before.


Background
Delayed gastric emptying (DGE) is a common postoperative complication following pancreatoduodenectomy (PD) that not only impedes the adjuvant therapy of the malnourished patient with pancreatic cancer but also increases the length of hospitalization.

Patients and Methods
surgical department of Evaggelismos general hospital. The selection of the procedure was based on whether the tumor invaded the duodenum or not. Mortality rate was about 16%. 26 patients received randomly a Witzel gastrostomy with gastropexy intraoperatively, which would be closed when the postoperative ileus subsided and the discharge from the gastrostomy tube was below 500 ml daily. Liquid diet (water, tea, and soup) would start on 5 pod and progressed depending on the patient tolerance. A gastrografin swallow was performed in addition to clinical assessment when the patients could not return to a standard diet by the end of the first postoperative week. If it would indicate slow gastric emptying, an opened pylorus, and an unobstructed antecolic gastrointestinal anastomosis, the gastrostomy tube was used as drainage. Upon tolerating solid diet, the patient would often be discharged with the gastrostomy tube closed and with instructions to appear after a week in our hospital to remove the tube.

Results
Pathologoanatomical examination of the 26 (11 female and 15 male) specimens ( Figure 1) after the operation demonstrated chronic pancreatitis in 3 patients, cancer of the ampulla of Vater in 4 patients, IPMN in 1, and cancer of the head of the pancreas in 17 patients. What should be highlighted is that the examination of all specimens revealed component of chronic pancreatitis and that perineural invasion was present in all specimens with pancreatic cancer. One female (30 years) had an echinococcal cyst in the head of the pancreas. 11 (5 female and 6 male) patients had a pylorus-preserving PD while the rest 15 (9 male and 6 female) a pylorus-resecting PD (Table 1).
Only four patients (15%), all of them with pylorusresecting PD, exhibited signs of gastroparesis. One of them presented leakage from the pancreaticojejunal anastomosis, while another one presented postoperative hemorrhage. The rest twenty-two patients had no postoperative gastric atony, even though four of them presented anastomotic leakage. The mean RBC units perioperatively transfused were 3 per patient.

Discussion
Over time, it has been assumed that DGE is strongly associated with the occurrence of other postoperative complications such as fistula formation, postoperative sepsis, and reoperation. Whether the duodenum is preserved or not, DGE occurs after both classic (with antrectomy) and pylorus-preserving pancreatoduodenectomy [34][35][36][37].
Several therapeutic alternatives for the treatment of DGE are reported in the literature like the usage of prokinetic drugs (motilin, erythromycin, and cisapride), postoperative cyclic enteral feeding, and several specifications for reconstructing the alimentary tract, preservation of vagal pyloric branches, vascular structures, and length of the duodenum.
Since gastric atony after these procedures is common (20-30%), in anticipation for prolonged requirement of gastric drainage, we regularly use a Witzel gastrostomy.
Indications to perform a gastrostomy, in historical order of their development, include the following: (1) to provide nutrition, (2) to allow retrograde dilatation of the esophagus, and (3) to provide postoperative decompression of the gastrointestinal tract [38]. Gilchrist (1953) was an early advocate of using a gastrostomy tube to decompress the stomach postoperatively, to avoid the complications of an indwelling nasogastric tube and to promote patient comfort [39]. Warshaw and Torchiana [40] found that patients after PPPD frequently required prolonged nasogastric intubation and could not tolerate a regular diet in the early postoperative period due to DGE. The Witzel technique [41], as described (Figures 2-4), has proved to be safe, reliable, and simple to perform. It offers the patient comfort (since a nasogastric tube is not necessary), immediate function for decompression, and low   [40,42]. Complications include internal migration, wound infection and peritonitis, when leakage occurs, or inability to remove it and dislodgement. The site of insertion of the gastrostomy tube is selected along the greater curvature of the body of the stomach, within approximately 2 or 3 cm of the gastroepiploic branches and on the anterior wall of the stomach. The selection of this site is based on two speculations.
Firstly, fixating gastrostomy on the anterior and middle portion of the stomach might not affect both gastric fundic lag phase and gastric antral motility, which are considered important for gastric emptying [31].
Secondly, fixating the stomach to the anterior wall keeps it taut enough to prevent torsion or angulation of the reconstructed GI anastomosis. Combination of gastropexy with an antecolic anastomosis seems to facilitate passive emptying of food into the jejunum with the help of gravity.
The gastrostomy placement provides through its "gastropexy effect" a slightly increased low esophageal sphincter (LES) pressure [43]. The speculation that it may slow gastric motility and lead to slower function of the GI anastomosis after PPPD [44] was not confirmed in our study.

Conclusion
To date, there is no reference in literature about the use of gastrostomy to prevent DGE. Our small series of patients showed that gastrostomy-gastropexy can be used, with safety and efficacy, as a preventive measure for DGE in patients undergoing pancreatoduodenectomy. Further studies and randomized trials are necessary in order to establish safe conclusions.

Data Availability
The data that support the findings of this case report are available from the corresponding author, G. Floros, upon reasonable request.

Conflicts of Interest
The authors declare that they have no conflicts of interest.