The type of pancreatic resection is based on the location of the tumor. Tumor of the head and uncinate process of the pancreas is managed with pancreaticoduodenectomy (PD), and tumor of the tail of the pancreas is managed with distal pancreatectomy (DP) with or without splenectomy. However, tumor of the pancreatic neck and body can be managed with viable alternative procedures, such as central pancreatectomy (CP) without oncological compromise [1]. With the increasing use of cross-sectional imaging, an increasing number of low-grade tumors are being diagnosed and managed [2]. The indolent nature of these low-grade pancreatic tumors and the low recurrence rate make CP an effective alternative to distal pancreatic resection [3]. The disadvantage of DP for low-grade pancreatic tumors is that the normally functioning pancreatic parenchyma is sacrificed with no added oncologic benefit. Dagradi and Serio performed the first central pancreatic resection and reconstruction in 1982 [4]. Since then, many surgeons have reported their experiences with CP [4]. The rate of postoperative pancreatic fistula (POPF) after CP is high because there are two cut surfaces and a small-diameter pancreatic duct in the distal pancreatic stump that need to be managed after CP [5,6]. Some reports have compared the outcomes of CP with those of DP [6]; hence, we wanted to compare the short- and long-term outcomes of CP with those of DP.

In 18 years (from 2003 to 2020), a total of 176 patients underwent surgery for low-grade pancreatic tumors. Of these, 67 patients underwent PD, and the remaining 109 patients underwent either CP (n = 37) or DP (n = 72). Eighty-eight patients (cases = 37, controls = 51) formed the unmatched study cohort after 21 patients were excluded based on the exclusion criteria. After propensity score matching, 74 patients (cases n = 37, controls n = 37) formed the matched cohort (Fig. 1).

Baseline clinical demographic parameters, such as age, sex, and preoperative comorbidities, such as diabetes, hypertension, ASA grade, and preoperative symptoms, were comparable between the two groups in both matched and unmatched study cohorts (Table 1). The duration of surgery, pancreatic duct diameter, and pancreatic texture were similar between the two groups in both matched and unmatched study cohorts.

The size of the lesion was significantly smaller in the case group in both matched and unmatched study cohorts (4 cm vs. 7 cm; p < 0.001), and the length of the residual pancreatic stump was significantly longer in the case group (5 cm vs. 4 cm; p < 0.001). It was found that blood loss was not significantly different in the unmatched study cohort, but it was significantly greater in the case group in the matched study cohort (Table 2).

The conventional approach for a neck or body lesion of the pancreas is either DP or PD [8]. This approach results in loss of normal functioning parenchyma, which can lead to insufficiency of pancreatic function with no added oncological benefit. Recent novel approaches for pancreatic parenchyma preservation, i.e., pancreatic enucleation, duodenum-sparing resection of the head of the pancreas, and CP, have become part of the surgeon’s arsenal, particularly for low-grade tumors [9].

A prerequisite for considering CP is that the benign nature of the lesion must be confirmed either preoperatively or intraoperatively. The presence of an invasive carcinoma during the final histopathological examination may necessitate reoperation [8]. A frozen section histological examination of the resected specimen was performed in all patients who underwent CP to rule out malignancy as well as negative margin status in the present study. Gao et al. [10] reported about CP in nine cases of ductal adenocarcinoma of the pancreas.

For large tumors and pancreatic tail tumors, DP may be the only surgical option available. We categorically excluded such patients from the present study to reduce selection bias. In a review of the literature, none of the other studies have specified such strict exclusion criteria.

In the last decade, more studies have reported about CP by minimally invasive surgery (laparoscopic or robotic) with acceptable morbidity and mortality [11,12]. However, all patients in the present study underwent open-technique CP.

After resection, one of the critical issues in CP is the management of two pancreatic stumps. The proximal stump can be managed by suture ligation, or division with a stapler or pancreatico-enteric anastomosis, falciform or omental patch reinforced suture closure [5]. The challenging aspect of CP is the management of the distal pancreatic stump, given the short distal stump length and the small size of the pancreatic duct. Usually, the distal pancreatic stump is managed by pancreatico-enteric anastomosis either by pancreaticojejunostomy (PJ) or pancreaticogastrostomy (PG) [13]. As a standard recommendation, stump less than 5 cm is a contraindication for CP. However, at our institute, we extended our criteria to perform CP even in patients with a 4-cm distal pancreatic stump length.

The reported POPF rates after PG and PJ after CP range from 0% to 36% and 0% to 63%, respectively [14]. A meta-analysis of studies on PD showed that PG was not superior to PJ with respect to POPF. However, in multicentre randomized controlled trials, the incidence of POPF was lower in patients undergoing PG than in patients undergoing PJ [15]. However, comparative studies between PG and PJ after CP have not been performed. In the present study, duct-to-mucosa PJ was preferred to dunking PJ when PD could be identified separately. Interestingly, Wayne et al. reported CP without anastomosis in 10 patients without POPF [14]. However, the length and status of the distal stump were not reported. Sauvanet et al. [16] also reported that the distal pancreatic stump was oversewn without an enteric anastomosis when the stump was small or atrophic. Recently, reports of end-to-end anastomosis of proximal and distal pancreatic stumps have been presented [17].

In the present study, the overall morbidity (Clavien-Dindo classification) or clinically significant morbidity (Clavien-Dindo grade 3 or greater) was similar between the two groups (unmatched study cohort 10.8% vs. 5.9%, p = 0.182; matched study cohort 10.8% vs. 5.4%, p = 0.158). In the CP group, two patients underwent pigtail catheter drainage of collection (Clavien-Dindo grade 3a) following POPF and one patient had multiple organ dysfunction (grade 4b). In the unmatched DP group, two patients (grade 3a) underwent pigtail catheter drainage and one patient developed single organ dysfunction (grade 4a). However, in the matched group, only two patients had grade 3a complication. The overall POPF rate was significantly higher in the CP group (unmatched study cohort 43.2% vs. 25.5%, p = 0.08; matched study cohort: 43.2% vs. 21.6%, p = 0.047). However, the rate of clinically relevant POPF was similar between the groups. However, a meta-analysis by Regmi et al. [18] (relative risk 1.64, p < 0.001) and Dragomir et al. [19] (odds ratio 2.24, p < 0.0001) showed a significantly higher incidence of clinically relevant fistulas after CP [17]. This observed difference might be due to our institutional practice of trans-jejunal external stenting of PD.

Other complications, such as PPH (unmatched study cohort 2.7% vs 0%, p = 0.238; matched study cohort 2.7% vs. 0%, p = 0.314), were comparable between the two groups in contrast to the previously reported series with a significantly higher risk of postoperative bleeding after CP [18]. Postoperative hospital stay was also similar between the two groups. These results could be due to lower rates of clinically relevant POPF after CP in our study. DGE, SSI, and pulmonary complications were comparable between the two groups. In the CP group, there was one case of postoperative mortality related to POPF.

Long-term outcomes, such as endocrine insufficiency, were significantly lower in the CP group, similar to the reported literature. However, exocrine insufficiency was similar between the two groups. The reported meta-analyses showed that CP was associated with significantly less exocrine insufficiency [18,19]. The incidence of postoperative exocrine insufficiency after CP varies widely, as it depends on the pre-existing pancreatic abnormality, the extent of the resection, and the presence of chronic pancreatitis [18]; and it also depends on the type of reconstruction (more in PG compared to PJ). In the CP group of our study, long-term outcomes, such as endocrine insufficiency (p = 0.532), exocrine insufficiency (p = 0.653), and weight gain (p = 0.395), after dunking and duct-to-mucosa anastomoses were similar between the two groups. As the sample size was small in the groups (dunking anastomosis, n = 18; duct-to-mucosa anastomosis, n = 19), it was difficult to determine which anastomosis could provide better long-term outcomes after CP.

Limited literature is available on the long-term quality of life after CP [4,19] The quality of life was similar between the CP and DP groups based on EORTC- quality of life [20]. The present study also compared the quality of life (SF-36 questionnaire) between the CP and DP groups. The general health and pain components of SF-36 were significantly better in the CP group. This could be explained by the statistically significant weight gain in the CP group at the long-term follow-up.

Our study is one of the largest series of CP resections in the subcontinent. However, the study has a few limitations. First, the retrospective design of the study and a large time span for sample acquisition may have introduced some confounders, leading to selection bias. Therefore, we used propensity score matching to reduce potential selection bias and confounding bias. Second, the size of the lesions could be one of the important factors for determining the surgical plan. The resulting sample size after including lesion size as one of the criteria for propensity score matching was small to compare the groups. Therefore, the size of the lesions was not included in propensity score matching. Third, no quantitative assessment criteria were used to diagnose exocrine insufficiency, which could have resulted in reporting bias. The major strength of the study is that, despite the small sample size, we included only those patients in the analysis who had been offered either CP or DP based on the surgeon’s discretion.

None of the patients in the present study had a recurrence during a median follow-up of 22 months (range, 11–118 months).

In conclusions, CP has a high morbidity rate, but the clinically significant morbidity and mortality rates are apparently similar when compared to those of DP. However, this needs to be validated by performing a larger sample study. CP has better endocrine sufficiency and similar exocrine sufficiency compared to DP. It is an oncologically non-inferior alternative procedure to DP with a good long-term quality of life, especially for low-grade pancreatic tumors.