Splenic vein resection together with the pancreatic parenchyma versus separated resection after isolation of the parenchyma during distal pancreatectomy (COSMOS-DP trial): study protocol for a randomised controlled trial

The aim of the COSMOS-DP trial is to establish the non-inferiority of the safety of resecting the splenic vein together with the pancreatic parenchyma compared with that of the conventional technique of isolating the vein from the pancreas before ligation and division during DP using mechanical staplers.

Patients undergoing open or laparoscopic DP for pancreatic body and tail cancer, intra-ductal papillary mucinous neoplasm, neuroendocrine tumours, mucinous cystic neoplasm, or metastatic pancreatic tumours or similar are eligible for inclusion in this study. In addition, simultaneous resection of the pancreatic parenchyma and splenic vein in one session will be rendered possible through the evaluation of preoperative imaging study findings. Patients indicated for spleen-preserving distal pancreatectomy (SPDP) will be excluded. The Warshaw operation (spleen-preserving and splenic artery/vein resection) will be included. The extent of lymph node dissection and whether the celiac axis is dissected will be left to the discretion of the surgeons. A detailed overview of all eligibility criteria is provided in Table 1.

This study is designed as a multicentre (45 institutes, Additional file 1: Table S1, Institution list) prospective randomised phase III trial; central randomisation and registration system will be applied (1:1). After assessing the patients for eligibility, they will be centrally randomised to either Arm A (separate resection of the splenic vein) or Arm B (combined resection of the splenic vein). Upon randomisation, the patients will be stratified by the surgical approach used (open or laparoscopic), institution and thickness of the pancreatic parenchyma (< 15 mm or ≥ 15 mm) [11], but not for the consistency of the pancreas. However, the consistency will be reported as one of the key variables in the case report form. We will use Pcock and Simon’s minimization method for random assignments and Mersenne Twister for random number generation (Fig. 1, flow diagram of the COSMOS-DP trial).

The primary endpoint is the incidence of PF grade B/C. The secondary endpoints will be outcome measures related to surgery, such as the operative time, volume of blood loss, preoperative thickness of the resected pancreatic parenchyma, haemostasis of the staple line, integrity of the staple line, incidence of pancreatic injury, need for additional sutures to securely close the pancreatic stump, duration of drainage tube placement, postoperative hospital stay duration, and incidence of conversion from laparoscopic surgery to open surgery. The outcome measures related to complications include the incidence of PFs of all grades, incidence of grade C PF, incidence of intra-abdominal haemorrhage, incidence of all complications, comparison of the thickness of the resected pancreatic parenchyma with the incidence of PF grade B/C, mortality, and incidence of thrombosis of the splenic vein (at one and six months after surgery).

This study was designed to prove the non-inferiority of Arm B compared with Arm A in terms of the primary endpoint [12].

This trial was designed to evaluate the non-inferiority of group B compared with group A in terms of the incidence of PF grade B/C. The incidence of clinically relevant PF (grade B/C) after DP using a stapler has been reported to be 1.9–20.3% in recent clinical trials [1, 2, 5, 13]. Therefore, for an assumed PF incidence rate of 10% with a non-inferiority margin of 9%, the difference of allowable PF incidence rate between groups A and B is 0.09, and the allowable odds ratio is 2.11. When the statistical analysis is performed for a significance level of α = 0.05 (one side) in a non-inferiority design, 138 patients are calculated to be required per arm, with a power 100 (1-β) of > 80%, under the assumption that a small number of patients may be deemed ineligible and may thus be excluded from the analysis. Furthermore, as approximately 5% of the patients are expected to be ineligible for surgery as indicated by the laparotomy or laparoscopic findings, the sample size was eventually increased to 304 patients (152 patients per arm).

As secondary endpoints, we will compare binary variables with Fisher’s exact test, continuous variables with the Mann–Whitney U test, and survival outcomes with the log-rank test. All results will be analysed using the full analysis set (FAS), which will include all patients except for those deemed ineligible after registration.

To evaluate postoperative PF, which is the primary endpoint, drainage tubes will be placed before closure of the abdomen. The number and sites of the drainage tubes that are inserted will be recorded on the data sheet. The timing of removal of the drainage tubes will also be specified.

To confirm that the surgical procedures are conducted as allocated at the time of central judgement, two photographs (before and after pancreatic transection) will be taken for all patients. Central judgement will be conducted biannually for all of the registered patients. At that time, the photographs will be reviewed by more than two members of the committee.

Antibiotics, plasma expanders, blood products, analgesic drugs, H2 blockers and proton pump inhibitors will be used for intra- and postoperative management at the discretion of the surgeons. In addition, there are no regulations on the drugs used to control complications and adverse events. Prophylactic administration of octreotide will not be permitted.

There are no regulations regarding the use of preoperative, intraoperative or postoperative chemotherapy or radiotherapy as treatments in this study. However, details including the treatment regimen and cycles given of the preoperative treatment are to be reported in the case report form.

The presence of thrombus in the splenic vein will be evaluated at one and six months after surgery by enhanced computed tomography or magnetic resonance imaging (Fig. 2, study calendar).

Interim analysis will be performed once, taking multiplicity into account using the Lan–DeMets method with O’Brien and Fleming type boundaries. The Data and Safety Monitoring Committee will independently review the interim analysis report and stop the trial early if necessary.

Central monitoring will be performed each year by the data centre to evaluate the study progress and ensure for study quality. The following aspects will be monitored: (1) data accumulation; (2) patient eligibility; (3) severe adverse events; (4) protocol deviations; (5) reasons for cessation or expiration of the protocol; (6) background factors of the patients; and (7) other problems concerning study progress and safety.

A complication is defined as an event occurring within six months after surgery.

The definition of PF is based on the International Study Group of Postoperative Pancreatic Fistula (ISGPF) classification [12].

The definition of delayed gastric emptying is based on the International Study Group of Pancreatic Surgeons (ISGPS) classification [14].

The definition of intra-abdominal bleeding is based on the International Study Group of Pancreatic Surgeons (ISGPS) classification [14].

The definition of other postoperative complication is based on the Clavien-Dindo classification [15].

The COSMOS-DP trial was opened in August 2016. At the time of submission for this paper (May 2018), protocol version is ver.2.2. The completion date is estimated to be December 2019.