After elective TP was first performed in the United States in 1942 , this procedure was applied to improve prognosis for advanced PDAC or prevent pancreatic fistulae after partial pancreatectomy , but the results of these attempts were disappointing. As a result, TP with curative intent for pancreatic tumor tends to be an uncommon operation for many surgeons. However, multifocal diseases such as IPMN remains an indication for TP, and considering recent refinements in diagnostic modalities and therapeutic interventions for primary tumors, surgeons seem likely to face patients requiring TP more frequently. Moreover, postoperative diabetes has been better managed by improvements in the rapid and long-acting insulin analogs and their combined use, and malnutrition status due to exocrine deficiency has also become manageable with high-quality pancreatic enzyme formula, resulting in better QOL. In this context, clinical cases that require TP involving remnant pancreatectomy may increase. In fact, in our hospital, 23 of 26 TP (88.4%) were performed in the second half of the study period, between 2009 and 2016.
Regarding indications for TP, our study revealed that the most frequent indication was multiple diseases; whether synchronous or metachronous diseases, 22 patients (22/26, 84.6%) underwent TP for multiple diseases. The most common multiple disease was IPMN-associated tumor such as IPMN, IPMC or PDAC concomitant with IPMN, as multifocal occurrence of IPMNs has been reported elsewhere [7–9]. In our case series, IPMN-associated tumors show a lower tumor grade, earlier stage at surgery and favorable prognosis, suggesting an underlying nature of low malignancy. The opposing cohort of patients with multiple diseases were those with de novo PDACs. De novo PDACs showed a higher tumor grade and more advanced stage at surgery, suggesting these tumors represent more aggressive tumors. The de novo PDAC group must represent one of the main multifocal pancreatic diseases, although fewer patients showed de novo PDACs (5/22, 22.7%) than IPMN-associated tumors. This lower frequency compared with IPMN-associated tumors could be partially explained by the fact that the prognosis of patients with PDAC after '‘initial surgery’' is poor, and improvement of treatment outcomes for the primary cancer could result in more cases of secondary PDAC representing multifocal disease. In fact, the cumulative 5-year incidence of second primary PDAC has been reported as 17.7%, markedly higher than that of metachronous gastric cancer . Moreover, comparison of patients with T-TP between the de novo PDAC and IPMN-associated tumor cohortsrevealed that the interval from initial surgery was shorter in the de novo PDAC group. These results suggested that surveillance for remnant pancreatic malignancy in the de novo PDAC group should be mandatory from the early period and at shorter intervals, along with monitoring for systemic recurrence. Ultimately, PDAC patients with multifocal potential will probably be identified through new genetic markers in the future, and if so, TP as a prophylactic surgery could achieve complete cure for patients in this group.
Although very few reports have analyzed the survival rate for secondary PDAC, Zhou et al  reported pooled analysis for 55 patients with secondary PDAC and the 5-year survival rate was 40.6%, higher than in patients with primary PDAC after initial pancreatectomy .
Conversely, the problem in the IPMN-associated tumor group is that both secondary PDAC and IPMC often recur after a considerably long interval, although these patients rarely develop recurrence outside the remnant pancreas, nor did any patients with non-invasive IPMC or IPMN, as reported . In our study, most secondary tumors in the IPMN-associated tumor group were found at an earlier stage, indicating that follow-up screening contributes to early detection for secondary tumors in this group.
In this study, comparative review between the O-TP and T-TP groups provided little information. One reason was that various diseases were mixed between both groups. With regard to the operative procedure, O-TP was naturally associated with longer operation time and greater bleeding compared to T-TP, although the extent of adhesion was another factor associated with surgical difficultyin some cases in the T-TP group.
Regarding postoperative nutritional status, glycemic control was not as difficult in our study as suggested in the past . In accordance with standard practice in our institution, all TP patients in this study were controlled by intensive insulin therapy in an attempt to mimic the normal pattern of insulin secretion, and to deliver replacement insulin using the concepts of basal and bolus insulin coverage. In addition to intensive insulin therapy, pancreatic enzyme formula in pancreatic enzyme replacement therapy (PERT) was indispensable, preventing fat maldigestion and malabsorption as well as deficiencies of fat-soluble vitamins as a consequence of exocrine pancreatic insufficiency [14,15]. Moreover, PERT reduces glycemic variability through the prevention of fatty stool . PERT combined with dietary management followed by an increase in insulin requirement can result in improvement of the nutrition status of patients.
The median HbA1c level of 8.0% in our patients was slightly high as in other reports [17–19], but at a level at which no life-threatening complications attributable to diabetes or hypoglycemia ordinarily occur . One limitation to our study was that the number of patients was small and some patients had not yet been followed for the long term, because TP remains a relatively rare procedure. Still, this study is valuable in that we focused on two entities of multifocal disease and because few reports have dealt with consecutive cases of TP from the same institution. Accumulation of additional cases is needed to definitively characterize the efficacy and long-term clinical outcomes of this surgical procedure.