Short- and long-term outcomes after pancreaticoduodenectomy following total gastorectomy: Report of case series and literature review.

PURPOSE: Pancreaticoduodenectomy (PD) following gastrectomy (TG) should be considered challenging even currently although its procedure and clinical value have been being standardized. Short- and long-term outcomes as well as standard reconstruction method following these procedures remain unclear. In order to clarify these issues, we reviewed worldwide English literature and 4 of our own cases of PD for patients with previous TG.
METHODS: Clinicopathological variables of 11 cases of PD for patients with previous TG were evaluated. Seven of these 11 were abstracted from a review of worldwide English literature and 4 of 11 were our own cases.
RESULTS: 3 cases was reconstructed using Y-limb made in previous TG and afferent loop syndrome (ALS) was observed in 2 of 3, in these cases whereas no cases of ALS were found in cases reconstructed using newly-made Y-limb. In cases where PD was indicated for cancer, early cancer death, defined as death related to cancer recurrence observed within 2 years after PD, was observed in 6 of 9 cases. Notably in cases of pancreatic cancer recurrent cancer was diagnosed within 1year after PD in 5 of 7 cases and 4 of these patients died of pancreatic cancer soon after recurrence.
CONCLUSION: In cases of PD following TG, previously-made Y-limb should not be used for reconstruction following PD because of probable cause of previously-made Y-limb for ALS. Long-term outcomes of PD after TG seemed unsatisfactory notably in cases of pancreatic cancer and thus application of PD for patients with previous TG should be carefully decided until reasonable explanation for this dismal outcome is obtained.

Introduction

Recently, there has been increasing the number of pancreaticoduodenectomy (PD) procedures in patients with previous gastrectomy. This is owed to the improved survival of patients who underwent total gastrectomy for gastric cancer [1], [2]. Although its procedure and clinical value have been being standardized, PD for patients who previously received total gastrectomy (TG) is considered challenging. Short- and long-term outcomes as well as standard reconstruction method following these procedures remain unclear.

In this study, we examined the patient characteristics and short- and long-term outcomes of PD following TG by reviewing our own single-center experience of 4 cases as well as cases previously reported in worldwide English literature.

Methods

We conducted a retrospective, descriptive analysis of PD in patients who previously underwent PD for various indications. Among 172 patients who underwent PD at our institute from 2001 to 2015, 4 patients had undergone TG before PD. Literature search revealed only 5 papers describing 7 cases of PD following TG. We reviewed 11 cases to collect the following data: age, sex, indication of TG, reconstruction method following TG, interval between TG and PD, indication of PD, reconstruction type of PD, operative time, bleedings, hospital days, complication, and prognosis. The median follow-up period among all cases was 10.5 months (range, 6–60).

Results

Characteristics of the patients undergoing TG

The demographics and operative findings of our 4 patients and additional 7 reviewed are summarized in Table 1. This summary includes 10 men and 1 woman with median age of 58 years (range, 45–68). Indication of TG was gastric cancer in 9 patients, gastric ulcer in 1 patient, and gastric malt lymphoma in 1 patient. Reconstruction method of TG was ‘Roux-en-Y’ in 9 patients, ‘esophagojejyunostomy’ in 1 patient, and ‘jejunal interposition' in 1 patient.

Table 1

List of patients who underwent pancreaticoduodenectomy after total gastrectomy.

Total gastrectomy					Pancreaticoduodenectomy
Case	Age	Gender	Indication	Anastomosis	Age	Interval (months)	Indication	Reconstruction	Operative time (min)	Blood loss (ml)	Complication	Rec,Y/N	Disease free (months)	Survival (months)	A/D	Ref.
1	58	M	gasrtic cancer	Roux-en-Y	69	132	pancreatic cancer	Roux-en-Y type	568	1009	afferent loop syndrome	Y	3	9	dead	†
2	54	M	gasrtic cancer	Roux-en-Y	78	288	IPMC	New- Roux-en-Y type	412	725	none	Y	9	12	dead	†
3	57	M	gastric malt lymphoma	Roux-en-Y	65	96	bile duct cancer	New- Roux-en-Y type	471	303	none	N	24	24	alive	†
4	51	F	gasrtic cancer	Jejunal Interposition	81	360	pancreatic cancer	New- Roux-en-Y type	409	513	none	Y	3	5.5	dead	†
5	65	M	gasrtic cancer	Roux-en-Y	69	43	pancreatic cancer	Roux-en-Y type	509	N/A	afferent loop syndrome	Y	N/A	N/A	dead	Ref. [4]
6	45	M	gastric ulcer	Roux-en-Y	46	15	pancreatic gastrinoma	New- Roux-en-Y type	503	N/A	none	N	N/A	N/A	alive	Ref. [4]
7	61	M	gasrtic cancer	Roux-en-Ya	71	120	pancreatic cancer	New- Roux-en-Y type	N/A	N/A	pancreatic fistula	Y	N/A	20	dead	Ref. [5]
8	64	M	gasrtic cancer	Roux-en-Y	68	48	bile duct cancer	Roux-en-Y type	445	1430	SSI	Y	15	24	dead	Ref. [6]
9	36	M	gasrtic cancer	esophagojejyunostomy with Braun	56	56	pancreatic cancer	New- Roux-en-Y type	672	938	none	Y	6	9	dead	Ref. [7]
10	62	M	recurrent gastric cancer	Roux-en-Y	72	120	pancreatic cancer	New- Roux-en-Y type	450	500	none	N	6	6	alive	Ref. [8]
11	68	M	gasrtic cancer	Roux-en-Y	80	144	IPMN	New- Roux-en-Y type	540	1300	none	N	6	6	alive	Ref. [8]

† our own case, N/A: Not avaiable.

total gastrectomy, distal pancreatectomy, and splenectomy.

Characteristics of the patients undergoing PD following TG

In the 11 patients, median age at PD was 69 years (range, 46–81), and median interval between TG and PD was 120 months (range, 15–360). Indication of PD was pancreatic ductal adenocarcinoma (PDAC) in 6 patients, lower bile duct cancer in 2, intraductal papillary mucinous carcinoma (IPMC) in 1, pancreatic gastorinoma in 1 and intraductal papillary mucinous neoplasm (IPMN) in 1.

Median operative time was 506 min (range, 412–672), with 509 min as median for those using same Y-limb made in previous TG (range, 445–568), and a median of 471 min (range, 409–672) for those with a new reconstructed Y-limb. The overall mean blood loss was 840 ml (range, 303–1430), with mean of 1220 ml loss in the group using same Y-limb (range, 1009–1430), and a mean of 713 ml (range, 303–1300) blood loss in those with newly constructed Y-limb. Three patients (27.3%) required blood transfusions. The median hospital stay of the patients was 30 days (range, 19–71). Four patients (36.4%) had postoperative complications including afferent loop syndrome (ALS) (2, 18.2%), pancreatic fistua (POF) (1, 9.1%), and wound infection (1, 9.1%). POF was grade A and was treated conservatively. Of 2 ALS cases, 1 case was our first case of PD following TG, and was required for reopation [3]. There was no operative mortality (Table 2). Three of 11 (27.3%) cases were reconstructed using Y-limb made in previous TG and ALS was observed in 2 of the 3 (66%) (Fig. 1a). The third patient in this group experienced SSI. Meanwhile, the group with a newly reconstructed Y-limb did not experience ALS (Fig. 1b) and only one patient suffered a postoperative complication in the form of pancreatic fistula.

Table 2

Operative findings and surgical complications.

Operative findings and surgical complications
Operative findings
Operation time (min)		506 (412–672)
Estimated blood loss (ml)		832 (303–1430)
Trasfusion (%)		3 (27.3%)
Length of hospital stay(day)		30 (19–71)
Surgical comlications
Clavien-Dindo Grade I	SSI	1 (9.1%)
Clavien-Dindo Grade II	Pancreatic fistula (Grade A)	1 (9.1%)
	Afferent loop syndrome	1 (9.1%)
Clavien-Dindo Grade IIIb	Afferent loop syndrome	1 (9.1%)
Total		4 (36.4%)

Fig. 1

Reconstruction methods after PD following TG.

a: The previously fashioned Y-limb was utilised for pancreatic and biliary reconstruction during the PD.

b: The afferent loop is staped and removed within the specimen, wheras a new Y-limb utilised for pancreatic and biliary reconstruction.

In cases where PD was indicated for cancer, early cancer death, defined as death related to cancer recurrence observed within 2 years (range, 5.5–24 months) after PD, was observed in 6 of 9 cases. Notably in cases of pancreatic cancer (PDAC/IPMC), recurrent cancer was diagnosed within 1 year (range, 3–9 months) after PD in 4 of 7 cases and 4 of these patients died of pancreatic cancer soon after recurrence (within one year).

Discussion

Pancreaticoduodenectomy is a technically demanding procedure, with a poor prognosis as well as poor survival benefit, ranging from 10 to 20% [9], [10], [11], [12], [13], [14]. It becomes particularly challenging in patients with a history of major abdominal surgery such as total gastrectomy. This additional difficulty in such operations is caused by the adhesions and anatomical complexity around the pancreas subsequent to the previous TG procedure. Due to the small number of reported cases who have gone through both these complex procedures, no consensus has been reached regarding the standard procedure for reconstruction following PD for patients with previous TG.

In our review of cases done in our centre and recorded in literature, the technique of Y-limb reconstruction during PD, was found to have an important impact on postoperative morbidity. ALS occured in patients whose the previously fashioned Y-limb was utilised for pancreatic and biliary reconstruction during the PD. In our centre, one case was performed with this technique, where resection of the duodenal part of the afferent loop fashioned in the TG was done, and using the remaining part of this loop for pancreatico- and hepaticojejunostomy (Fig. 2). We applied this procedure to reduce the number of intestinal anastomoses. As a result, however, the jejunal limb used was markedly short, resulting in the development of refractory cholangitis due to ALS [3]. Besides avoiding ALS, the cases where we reconstructed a new Y-limb had less median operative time (471 min) and mean intraoperative blood loss (713 ml) as compared to cases where same Y-limb fashioned in TG was used.

Fig. 2

Hepaticojejunostomy was performed using same previous afferent loop after total gastrectomy. The afferent loop was only 15–20 cm long.

Similarly, cases from other centres reported ALS occuring in one case out of two (50%) where the jejunal limb made in a previous TG was utilised in the PD. The rest of the cases in our centre and in other centres (n = 8), where a new Y-limb was refashioned during PD, no incidence of ALS was recorded.

Therefore, although number of cases is small, it may be safe to consider the reconstruction of a new jejunal Y-limb during PD in patients with previous TG. This is thought to avoid intestinal kinking and excess tension on the anastomosis, especially when the jejunal limb is short, and in turn avoid postoperative development of ALS.

The long term outcome, on the other hand, seems to be poor for patients who have pathologically proven pancreatic cancer (n = 7). Notably in cases of pancreatic cancer (PDCA/IPMC), recurrent cancer was diagnosed within 1 year after PD in 4 of 7 cases. Those 4 cases died of pancreatic cancer shortly after the diagnosis of recurrence. One out of the other 3 cases died within 2 years. A questionnaire was sent to the authors of the other reported cases to clarify the cause of death, but we did not get a clear answer.

Although this seems comparable to the long term survival of PD patients (10–20% 5 year survival), but it demands further evaluation of larger patients series who underwent similar procedures, in order to evaluate the feasibility of such complex procedure. Meanwhile, the possibility of having an immunological and/or oncological disadvantage by having a previous TG must be investigated.

In conclusion, PD should be attempted in TG patients, but with attention to the refashioning a new Y limb. However, because of the poor prognosis of PD following TG for pancreatic cancer, the indication for PD following TG should be carefully considered while taking the decision.

Authors’ contributions

Satoshi Yokoyama made the conception and design of this case report. Authors other than SA contributed to the collection, analysis, and interpretation of the data. SA wrote the draft manuscript, and other authors performed the critical revision of the manuscript. All authors gave final approval of the version to be published. SA has overall responsibility and guarantees the scientific integrity.

Competing interests

The authors declare that they have no competing interests.

Consent for publication

Written informed consent was obtained from the patient for the publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.