Experience on postoperative complications of laparoscopic pancreaticoduodenectomy

Abstract

Objective: To summarize the incidence and characteristics of postoperative complications after laparoscopic pancreaticoduodenectomy(LPD), and to share our experience on management of complications. Methods: The clinical data of 320 LPD performed by a single team in Sir Run Run Shaw Hospital and Zhejiang Provincial People's Hospital between September 2012 and September 2017 were retrospectively analyzed, among which there were 196 males and 124 females with age of (60.2±11.6) years old.There were 306 patients who underwent standard LPD, and 14 patients who underwent extended LPD. The patients were divided into 2 groups of former 160 LPD and later 160 LPD according to the time order. By analyzing the differences of clinical outcomes between the two groups, especially focusing on the incidence of postoperative complications.The experience on management of complications was concluded. The prior surgical history of latter group was significantly higher than the former group(30.0%(48/160)vs. 18.8%(30/160), χ(2)=5.49, P=0.019), and the rest of baseline characteristics remained the comparable (P>0.05). For resectable lesions, LPD was performed by "No back" approach, following the principle of "From distal to cephalad, from ventral to dorsal, and from left to right" . As for the borderline resectable patients, LPD was performed by "Easy first" strategy. Student t test, χ(2) test or Fisher test was used to analyzed the data between the two groups respectively. Results: Of 320 LPD patients, 306 cases underwent standard LPD, 14 cases underwent LPD with resection of other organs.There were 278 LPD cases who followed "No back" approach, and 42 cases who followed "Easy first" strategy because of difficulty in creating the retro-pancreatic tunnel. And the overall morbidity was 32.2%(103/320) with reoperation rate of 5.3%(17/320). The perioperative mortality was 0.6%(2/320). The operation time of latter group was ((346.6±48.8)minutes), which was shorter than that of former group((358.0±54.4)minutes)(t=1.97, P=0.048). The blood loss of former and latter group remained comparable((207.9±135.8)ml vs.(189.6±121.4)ml, P=0.205). However, in subgroup analysis, the patients with blood loss less <200 ml of latter group decreased significantly from 59.4%(95/160) to 47.5%(76/160)(χ(2)=4.53, P=0.033). The overall morbidity of latter group was 28.8%(46/160), indicated a decrease from 35.6%(57/160) of former group without significant difference(P=0.188). Moreover, Grade A/B/C pancreatic fistula rate, Grade A/C bile leakage rate, Grade B/C postoperative hemorrhage rate of the later group tended to decrease, although they also didn't reach a significant difference. However, the abdominal infection rate decreased significantly(χ(2)=3.93, P=0.047). The length of hospital stay remained comparable(P=0.156). Conclusions: The most common complications after LPD were postoperative hemorrhage and pancreatic fistula. With specialized team and accumulated experience, the morbidity can decrease progressively by analyzing the leading cause and improving the technical skills.