This is a retrospective clinical study on a collected data base. Between March 2016 and December 2018, a total of 102 patients who underwent LPD were enrolled in this study according to the inclusion criteria: (1) tumors of pancreas, duodenum and lower bile duct; (2) biliary obstruction caused by benign tumors (3) all patients had complete laboratory test data; (4) no anticancer treatments before the operation; (5) curative resection of all tumor nodules; and (6) complete patient records and follow-up data. Patients who had tumour metastasis and other organ resection were not considered eligible for this study. These surgeries were performed by Dr. De-Wei Li and Dr. Xiong Yan, who are experienced laparoscopic surgical experts and have completed more than 50 cases of LPD. Among these cases, 34 patients were prospectively included in enhanced recovery (ERAS group). These patients were compared with 68 consecutive patients after LPD under the care of the same consultants previously receiving standard care during an equal time frame (non-ERCP group, non-ER, Table 1). The Clavien-Dindo classification was used to evaluate the postoperative complications (10). This study protocol conformed to the ethical guidelines of the 1975 Helsinki Declaration and has been reviewed and approved by the Ethics Review Board of the First Affiliated Hospital of Chongqing Medical University (2022–031). All patients and/or their legal guardians were required to provide informed consent to participate in this study.

ERAS protocols were designed based on previously published ERAS guidelines (1), and we further refined and established our own standardized programmes (Table 2). Early oral intake and ambulation, active pain control, avoidance of unnecessary indwelling medical tubes, and promotion of patient autonomy were encouraged if the condition of the patients allowed. After surgery, patients went go to the observation unit and then moved back to the ward. If vital signs were unstable, the patients went to ICU postoperatively. On postoperative day (POD) 1, anti-thrombotic prophylaxis, especially VTE prophylaxis and post VTE prophylaxis were performed. From POD 1, the patients began ambulation such as mobilization out of bed for >1 h or at least 3 times. If necessary, ambulation assist equipment could be used. Appropriate glucose loading was used on the day of surgery based on the energy need of the patients. Postoperative oral intake of liquids was started on POD 1–2, and elemental diet was started on POD 2, and solids were started on POD 3–4. Intake of rice porridge was started on POD 3 in parallel with elemental diet. Postoperative fluid management was adjusted based on oral intake. The intravenous fluid infusion volume was controlled within 1000–1500 ml. Nasogastric tubes were removed on POD 1 if nasogastric output was <300 ml/day. The urinary catheter was also removed on POD 1. Intra-abdominal drains were removed on POD 3 if amylase was less than 3-fold higher than the serum level and output was <50 ml/d. Physiotherapy was started on POD 1 and continued until discharge. For postoperative prophylaxis, piperacillin/tazobactam sodium was administered intravenously on the operative day (4.5 g, three times). The compliance with ERAS and non-ERAS core elements of each patient was recorded as either 0 for non-accomplished or 1 for accomplished.

The standard surgical procedures were performed in all patients, including gastrointestinal anastomosis and the pancreaticojejunostomy (anastomosis of pancreatic duct mucosa and jejunum mucosa). The intra-abdominal drains were routinely placed at the site of the bilioenteric anastomosis and pancreaticojejunostomy.

SPSS 24.0 statistical software was used to analyse the data. The continuous variable data are shown as the mean ± standard deviation. Continuous variables were compared using Student's t-tests with a normal distribution or nonparametric Mann–Whitney U-tests with an irregular distribution. Count data were analysed using x² test, Fisher's exact test and the rank sum test. A P value <0.05 was considered statistically significant.

Clinical characteristics between two groups were compared using the following risk factors: gender, age, body mass index (BMI), American Association of Anaesthesiologists (ASA) score, and Child-Pugh score as well as measurements of white blood cells (WBC), neutrophils, haemoglobin, albumin, total bilirubin (TB), direct bilirubin (DB), alanine aminotransferase (ALT), and aspartate aminotransferase (AST) (Table 1). This study included 58 men (56.9%) and 44 women (43.1%) with an average age of 60.5 ± 9.2 years. The two groups were similar in terms of demographics and surgical characteristics (all P > 0.05). There was no difference in mean age, BMI, ASA score and Child-Pugh score between the ER and non-ER groups (P > 0.05). Common laboratory tests, including WBC, TB, DB, AST, ALT and albumin, did not differ significantly between the two groups (P > 0.05).

There were no statistically significant differences were observed between the ER and non-ER groups in duration of surgery, intraoperative blood loss and transfusion, which are presented in Table 3. In the ER group, the mean duration of surgery was 462.7 (300–720) min, the operative blood loss was 523.5 (250–1200) ml, and 8.8% of patients received an intraoperative transfusion. However, in the non-ER group, the mean duration of surgery, operative blood loss and proportion of patients who received an intraoperative transfusion were 450.9 (300–760) min (P = 0.627), 537.5 (250–1500) ml (P = 0.800) and 11.8% (P = 0.746), respectively. The intravenous fluid infusion rate was 4.0 ml/kg/h. Usually, the ratio of crystalloid to colloid was from 2:1 to 3:1. The ER group had significantly lower total fluid and crystalloid volume than non-ER group (P < 0.01). There was no significant difference in the colloid volume between two groups. All patients used analgesic infusion pump in both groups.

The complications included pancreatic or biliary fistula, haemorrhage, abdominal or pulmonary infection, unplanned reoperation, mortality rate and readmission within 30 days of the operation (Table 4). The overall complication rate was 32.4% (11 of 34 patients) in the ER group compared with 35.3% (24 of 68) in the non-ER group (P = 0.281). According to the clinical effect, postoperative pancreatic fistula has three grades (grade A, B and C) based primarily on the International Study Group on Pancreatic Fistula (ISGPF) (11). There were 7 and 14 patients with a grade A pancreatic fistula happened in the ER group and non-ER group, respectively. 3 and 5 patients in the ER group and non-ER groups developed a grade B pancreatic fistula after surgery, respectively. These patients were healed after treatment with abdominal puncture and B-ultrasound guided-drainage. There was no significant difference in the pancreatic fistula ratio between the two groups (P = 0.903). There was one patient with biliary fistula, delayed gastric emptying and pleural effusions or ascites in each group (P = 0.615), respectively. One patient with postoperative haemorrhage in non-ER group underwent a second surgery. Two patients with postoperative haemorrhage in ER group received blood transfusion and hemostasis treatments other than surgery (P = 0.216). In ER group, there was no patient with abdominal, pulmonary or wound infection, thrombosis, unplanned reoperation and dead. Correspondingly in patients with non-ER, there was one patient producing abdominal, or wound infection, thrombosis and unplanned reoperation, respectively (P = 0.480). One patient was dead (P = 0.480), and three patients underwent pulmonary infection (P = 0.216). Except one unplanned reoperation, the other postoperative complications were treated conservatively. One patient with ER chose readmission in 30 days due to ascites. According to the Clavien–Dindo grade (10), there was no significant difference in grade ≥ 3 complications between ER group and non-ER group (2 vs. 5 patients; P = 0.783). There was no difference in postoperative complications between the patients after LPD performed by Dr. Li and Dr. Yan. The median duration of postoperative fluid management was 9.5 days (range: 7.5–14 days) in the ER group and 14 days (range: 12.5–22 days) in the non-ER group (P < 0.01).

The mean total medical costs per patient were $2.1 ± 0.7 × 10⁴ and $2.3 ± 0.7 × 10⁴ in ER group were lower than those in non-ER group. in the ER and non-ER groups, respectively (P = 0.017). Postoperative mean costs per patient were $2.0 × 10⁴ in the ER group and $2.2 × 10⁴ in the non-ER group (P = 0.019). ERAS was significantly associated with lower costs for medication, housing, nursing care and disposable material costs (P < 0.05) in the ER group compared with the non-ER group. The surgical and anesthetic costs were similar in the two groups (P > 0.05). In the enhanced recovery group, these patients had some extra radiological examinations or laboratory examinations to prevent accidental complications due to the enhanced recovery. Other laboratory and radiologic examination costs had no significant difference (P > 0.05).

The recovery indicators for patients with ERAS or non-ER were comparatively analysed. The retention time of nasogastric tubes, urinary catheters and intra-abdominal drains in the ER group were significantly shorter than that in the non-ER group (P < 0.001). In addition, oral elemental diet and targeted mobilization after surgery in the ER group also occurred much earlier than that those in the non-ER group (P < 0.001) (Table 5). In addition, there were several patients deviating from the ERAS protocols. The nasogastric tube was removed on POD 1 in 26 (76.5%) patients, but four patients (11.8%) could not start oral intake of an elemental diet on POD 2. The overall compliance rates to protocol were 89.2% in ER group and 57.4.8% in non-ER group (P < 0.01), respectively. As shown in Table 5, early removal of urinary catheter and intra-abdominal drains were more difficult to follow in both groups. Compare with ER group, it was not easy to perfome vomiting prophylaxis, early removal of nasogastric tube and early oral elemental diet in patients with non-ER (P < 0.001).