Duodenum-Preserving Resection of the Pancreatic Head versus Pancreaticoduodenectomy for Treatment of Chronic Pancreatitis with Enlargement of the Pancreatic Head: Systematic Review and Meta-Analysis

The results of this meta-analysis show that DPPHR should be established as first-line treatment because of lower level of severe early postoperative complications, maintenance of endocrine pancreatic functions, shortening of postoperative hospitalization time, and increase of quality of life compared to pancreaticoduodenectomy.


Introduction
Chronic pancreatitis (CP) is defined as "A continuous inflammatory process causing permanent structural damage to the pancreatic gland, which ultimately results in impairment of the gland's exocrine and endocrine function" [1]. About 30% of CP patients have inflammatory enlargement of the pancreatic head (PH) [2]. PH enlargement can result in stenosis of the common bile duct or obstruction of the main pancreatic duct [3]. Hence, patients with inflammatory enlargement of the PH usually require PH resection to resolve these complications. The traditional Whipple procedure is first-line therapy for this type of CP, but it is associated with a high prevalence of morbidity and mortality and reduced quality of life (QoL) [4,5]. These disadvantages are attributed to extensive resection, including the removal of the duodenum and a large portion of the pancreas. The duodenum plays an important part in the regulation of digestive processes. In 1990, Beger and Buchler introduced duodenum-preserving pancreatic head resection (DPPHR) to preserve the duodenum and limit resection of pancreatic tissue [6]. A similar procedure was proposed by Ho et al. in which local resection of the PH with longitudinal pancreaticojejunostomy is carried out [7]. The literature suggests that both of these procedures are suitable for the treatment of this type of CP. Several studies have discussed the potential superiority of DPPHR over other surgical methods, but a systematic and quantitative review summarizing the available evidence is lacking. Therefore, we undertook a meta-analysis to compare the effectiveness and safety of DPPHR versus pancreaticoduodenectomy (PD) or pyloruspreserving pancreaticoduodenectomy (PPPD) for CP.

Methods
We searched for journal articles published from January 1990 to September 2016 by electronic and manual means. We searched the databases of PubMed, the Cochrane Library, Web of Science, and EMBASE using the following search terms: (pancreaticoduodenectomy OR duodenopancreatectomy OR pancreatectomy OR pylorus-preserving OR Whipple or PD or PPPD) AND (duodenum-preserving pancreatic head resection OR duodenum-preserving OR Beger or longitudinal pancreaticojejunostomy OR longitudinal pancreaticojejunoanastomosis OR LPJ OR Frey OR Beger OR DPPHR) AND (chronic pancreatitis). A language restriction was not applied and the search was carried out by two independent investigators ( Figure 1).

Inclusion Criteria.
We searched for studies comparing DPPHR with PD or PPPD for CP with inflammatory enlargement of the PH. We also searched for studies in which elective surgery was planned for patients diagnosed with CP in the PH. If the same study had been published more than once then the latest publication was used.

Exclusion Criteria.
A study was excluded if (i) the method of surgery was not reported; (ii) a comparison between the DPPHR group with a PD group or PPPD group was not made; (iii) the prevalence of postoperative complications and mortality as study outcomes was not reported; (iv) the study had been reported before; (v) it had design flaws and was of low quality; (vi) it was an abstract, case report, letter, comment, or review without original data; (vii) if it presented insufficient data.
2.3. Literature Screening. All literature was screened by two independent investigators. If the two investigators disagreed, then they tried to resolve the disagreement through discussion. If discussion failed, the final decision was made by a third investigator. EndNote reference management software was used to search and remove duplicate studies.  Table 1.

Data
The quality of RCTs was evaluated based on the Jadad scale, which was used to assess randomization, concealment of allocation, blinding, and withdrawals in each study. Each item was given a score of 0-2, and the maximum total score was 7. If the total score was ≥4, the RCT was of "high" quality. Observational clinical studies (OCS) were scored based on the Newcastle-Ottawa system, which involves assessment of selection, comparability, and exposure/outcome. The maximum total score was 9. If the total score was ≥7, the OCS was considered to be of "high" quality.

Assessment of the Risk of Bias of RCTs.
For the included RCTs, assessment of the bias risk involved six parameters: allocation concealment, incomplete outcome data, blinding, selective reporting bias, sequence generation, and other potential sources of bias. Assessment was based on a quality checklist recommended in the Cochrane Handbook. "Yes" indicated a "low" risk of bias; "unclear," an "uncertain" risk of bias; "no," a "high" risk of bias ( Figure 2).

Incidence of Pancreatic Fistula.
Eight included studies reported the incidence of postoperative pancreatic fistula; we pooled data from the eight studies to compare DPPHR group with PD/PPPD group. The results of meta-analysis show that there is no difference between two groups in the rate of POPF (OR = 0.76; 95% CI, 0.35-1.69; = 0.51; 2 = 0% for heterogeneity). Therefore, using a fixed model. The metaanalysis of RCTs and OCS subgroup reveals no heterogeneity among studies and the incidence of POPF was not statically different between two groups. [RCTs ( 2 = 0%, OR = 0.83;

Incidence of Wound Infection.
Four included studies reported the incidence of wound infection after PD/PPPD and DPPHR. 2 ( 2 = 40%) revealed no obvious heterogeneity among these studies; therefore using a fixed model, there was no significant difference between DPPHR and PD/PPPD group in the incidence of wound infection (OR = 1.00; 95% CI, 0.46-2.17; = 0.99) ( Figure 5).

Incidence of Endocrine Insufficiency.
Eight included studies reported the incidence of postoperative endocrine insufficiency. In these included studies, endocrine pancreatic functions were estimated at least 6 months after operation. We pooled data from the eight studies to compare DPPHR group with PD/PPPD group. The results of meta-analysis show that there is statistical difference between two groups in the rate of postoperative endocrine insufficiency (OR = 0.35; 95% CI, 0.21-0.61; = 0.0002; 2 = 35% for heterogeneity). Therefore, using a fixed model. The meta-analysis of RCTs and OCS subgroup reveals the incidence of postoperative endocrine insufficiency was statically different between two groups [RCTs ( 2 = 0%, OR = 0.19; 95% CI, 0.08-0.50;  RCT: randomized controlled trial, PD/PPPD: pancreaticoduodenectomy, and DPPHR: duodenum-preserving pancreatic head resection. Jadad scale system: The Jadad scale, sometimes known as Jadad scoring or the Oxford quality scoring system, is a procedure to independently assess the methodological quality of a clinical trial. It is the most widely used assessment in the world. The Newcastle-Ottawa System: the quality of the nonrandomized studies was assessed by using this system. The quality of the studies was evaluated by examining three items: patient selection, comparability of groups, and assessment of outcome. Test for overall e ect: Z = 0.46 (P = 0.64)

Incidence of Exocrine Insufficiency.
Five included studies reported the rate of exocrine insufficiency after PD/PPPD and DPPHR. Pancreatic exocrine insufficiency was defined as diarrhea and steatorrhea, which improved with pancreatic enzyme replacement. 2 ( 2 = 65%) revealed obvious heterogeneity among these studies; therefore random model was applied. There was no significant difference between DPPHR and PD/PPPD group in exocrine insufficiency (OR = 0.43; 95% CI, 0.12-1.47; = 0.18) (Figure 7).

Postoperative Weight Gain.
Five included studies reported the rate of weight gain after PD/PPPD and DPPHR. In these included studies, the rate of weight gain was estimated at least 6 months after operation. 2 revealed no obvious heterogeneity among these studies. Therefore, using a fixed model, there was significant difference between DPPHR and PD/PPPD group in postoperative weight gain (OR = 5.85; 95% CI, 3.27-10.45; < 0.00001) (Figure 8).

Postoperative Functioning Scale Scores.
Another objective criterion in addition to the above basic parameters of a specific surgical method is currently global improvement of the quality of life. The objective outcome assessment of surgical treatment has been made by the EORTC QLQ, used for patients with CP; the system of EORTC QLQ-C-30 includes five terms, such as physical status, working ability, emotional functioning, cognitive functioning, and Büchler et al.

Subgroup
Analysis. The procedures of duodenumpreserving pancreatic head resection included the Beger and Frey two methods of surgery. Beger procedure was performed with jejunal Roux-en-Y loop which drained the residual pancreas via an end-to-end or end-to-side pancreaticojejunostomy to the body of the pancreas and a side-to-side pancreaticojejunostomy to the excavated pancreatic head remnant and opening intrapancreatic choledochus and suturing the incision to bilateral pancreas tissues. Frey procedure included local resection of the enlarged pancreatic head and a longitudinal incision of the dilated duct; reconstruction was accomplished by longitudinal pancreaticojejunostomy and the common bile duct was drained by cholangiojejunostomy. In order to reduce the heterogeneity resulting from different   Test for overall e ect: Z = 2.52 (P = 0.01) Heterogeneity:  2 = 50.45;  2 = 6.57, ＞＠ = 3 (P = 0.09); I 2 = 54%     ; therefore the fixed effects model was used (Figure 14). ; therefore the fixed effects model was used (Figure 15).

Publication Bias.
Funnel plots were created to assess the publication bias in our meta-analysis of included studies. In the absence of publication bias, it assumes that studies with high precision will be plotted near the average, and studies with low precision will be spread evenly on both sides of the average, creating a roughly funnel-shaped distribution. Deviation from this shape can indicate publication bias. There Keck   was no evident asymmetry in the funnel plots (Figures 16 and  17), suggesting a low probability of publication bias.

Discussion
CP requires conservative treatment [23]. Complications such as bile duct stenosis as well as duodenal, pancreatic duct, or vascular obstruction often necessitate surgical intervention for CP with PH enlargement. Elective surgical procedures undertaken in CP patients can be divided into resection or drainage procedures. The Whipple procedure is first-line therapy for PH tumors. To ascertain if preservation of the duodenum benefits patients, some randomized studies have been done to compare the morbidity, mortality, pain relief, and exocrine/endocrine function between these two surgical approaches.  The result of our meta-analysis showed that PD/PPPD and DPPHR were not significantly different with regard to pain relief, pancreatic fistulae, infection, or postoperative mortality. In CP, pain mechanisms in patients with chronic pancreatitis are incompletely understood and probably multifactorial. Many factors, such as pancreatic duct obstruction, neuropathic changes, alterations in nociception, maybe link to pancreatic pain. Conservative and endoscopic therapy will have less benefit in pain relief, while DPPHR and PD/PPPD are all effective treatments to relieve pain [24]. DPPHR and PD/PPPD were equally effective in controlling pain and had an acceptable low mortality rate. The main purpose of DPPHR is to preserve the integrity of the digestive tracts.
Our results revealed DPPHR to have a low prevalence of delayed gastric emptying compared with PD. We hypothesize that preservation of continuity of the gastroduodenal passage is important for reducing the risk of delayed gastric emptying. DPPHR could also reduce the duration of postoperative hospitalization significantly and reduce the expense of treatment. The longer length of stay in PD was likely attributable to the higher incidence of delayed gastric emptying. Moreover, our group meta-analysis and subgroup meta-analysis revealed that DPPHR could lower the prevalence of endocrine inefficiency compared with PD, whereas there was no significant difference between DPPHR and PD in the prevalence of exocrine inefficiency. Approximately 40-60% of the pancreas is resected in PD compared with 10-40% in DPPHR, so the latter can preserve more normal pancreatic tissue and protect exocrine and pancreatic functions. Compared with PD, DPPHR could increase postoperative weight gain significantly. Indeed, some researchers have reported that 80-90% of patients can increase weight after DPPHR.
Surgical treatment for CP can lower the risk of pain and complications, but also improve the quality of life, physical status, and social and occupational rehabilitation. The European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire (QLQ) is a suitable and reliable tool for assessing the global QoL of CP patients [25]. The EORTC QLQ-C-30 comprises five terms: working ability, physical status, cognitive function, emotional function, and social function. Our meta-analysis showed that DPPHR had more advantages than PD in terms of improving QoL. Compared with PD, DPPHR was associated with less damage to the retroperitoneal nerve plexus. Moreover, DPPHR preserves the secretion function of the antrum and duodenum, which benefits postoperative recovery. Our meta-analysis suggests that DPPHR should be adopted as a new standard procedure in the treatment of pancreatic head complications in chronic pancreatitis.
A major limitation of our meta-analysis was that only a small number of high-quality RCTs was included. The surgical experience and methods used at different hospitals and specialist centers could have produced different outcomes and increased the heterogeneity between the included studies. Also, the treatment of complications may have affected the outcome of the RCTs and OCS included in this metaanalysis.

Conclusions
Our meta-analysis revealed that DPPHR was more beneficial than PD/PPPD in reducing the prevalence of delayed gastric emptying; endocrine insufficiency; duration of postoperative hospitalization. Also, DPPHR increased the QoL of patients. No significant differences were found with regard to the prevalence of pain relief, pancreatic fistulae, wound infection, exocrine insufficiency, or mortality between the two approaches. A similar prevalence of mortality for DPPHR and PD/PPPD was not surprising given their comparable degree of surgical complexity. Therefore, DPPHR should be established as first-line treatment because of lower level of severe early postoperative complications, maintenance of endocrine pancreatic functions, shortening of postoperative hospitalization time, and increase of quality of life compared to pancreaticoduodenectomy.