Surgical resection remains the most primary curative option for patients with gastric cancer (GC) [1]. Laparoscopic surgery for the treatment of GC is becoming an increasingly popular alternative for early-stage disease [2, 3] to advanced lesions [4, 5], and this approach not only has short-term advantages, such as fast recovery of bowel function, obvious reduction in postoperative pain, and considerable shortening of the length of hospital stay [6, 7], but also comparable long-term oncologic outcomes to open surgery [8, 9]. Improving the safety of laparoscopic procedures to the greatest extent remains a primary issue for surgeons. Perioperative early unplanned reoperation (EURO, within 30 days of hospitalization) is usually aimed at resolving severe or life-threatening postoperative complications. On the one hand, EURO seriously affects postoperative recovery, obviously adds an economic burden and can even lead to early death in some patients [10,11,12]. On the other hand, EURO greatly increases the medical workload of healthcare workers. In addition to postoperative morbidity and mortality, EURO has attracted the attention of more surgeons recently and has become widely accepted as an important indicator of the quality of medical care [13,14,15].

Understanding the main causes and time intervals of EURO among GC patients undergoing gastrectomy, and in particular, evaluating the risk factors for EURO before surgery, might help surgeons to reduce the incidence of EURO and improve the safety of surgery. Most previously published studies have primarily considered the presence of morbidity and mortality after gastrectomy for GC [6, 7, 16] without focusing on the severity and urgency of EURO, making the conclusions less objective. Data directly comparing differences in EUROs following initial laparoscopic-assisted gastrectomy (LAG) and open gastrectomy (OG) have been quite limited so far [6, 7, 10, 15]; in particular, detailed data regarding perioperative consequences, treatment strategies, and high-risk factors for EURO following LAG and OG are still not available. Therefore, we investigated the differences in the incidence, causes, intervals, and clinical outcomes of EUROs performed in patients who underwent LAG or OG. The results of the current study may contribute to providing valuable clues for the prevention and treatment of EURO by precisely identifying perioperative consequences and high-risk factors according to different surgical approaches.

Materials and methods

Patients and methods

Between January 2005 and December 2014, 4124 patients with primary GC underwent radical gastrectomy at the Department of Gastric Surgery, Fujian Medical University Union Hospital, Fuzhou, China. A retrospective analysis was performed using a prospectively maintained comprehensive GC database [17]. The indications for both OG and LAG in this study included a preoperative diagnosis of clinical T1–T4a (cT1–T4a)-stage primary gastric adenocarcinoma; patients with remnant GC, T4b-stage lesions or distant metastatic tumors and those requiring emergency surgery from bleeding or perforation were excluded. Patients were informed of the potential complications of the procedure and the advantages and disadvantages of a laparoscopic vs. an open approach, and written informed consent was obtained from all the patients prior to surgery. In this study, 2608 patients undergoing LAG and 1516 patients undergoing OG were eligible for enrollment. This study was approved by the Institutional Review Board of Fujian Medical University Union Hospital (Approval number: 20070428).

Lymphadenectomies were classified based on the 3rd English editions of the Japanese Classification of Gastric Carcinoma [18]. The data on the extent of lymphadenectomy were revised retrospectively according to a recently published version of those guidelines [18]. The tumors were staged according to the final pathological results based on the 7th Edition of the American Joint Committee on Cancer (AJCC) staging system [19].

Variables and definitions

The potential risk factors for EURO were extracted from the database, along with age, gender, comorbidity, American Society of Anesthesiologists (ASA) score, body mass index (BMI), preoperative malnourishment, neoadjuvant treatment, abdominal surgery history, tumor location, tumor size, pathological TNM stage, LAG or OG, gastrectomy type, reconstruction type, operative time, intraoperative blood loss, and number of resected lymph nodes (LNs). Postoperative morbidity was defined as any one or more of the following: abdominal bleeding, anastomotic bleeding, anastomotic leakage, abdominal infection, ileus or mechanical obstruction, pancreatic fistula, incisional infection, incision dehiscence, and pneumonia. Postoperative morbidity was graded based on the revised version of the Clavien–Dindo classification system suggested by Dindo et al. [20] For patients with more than one complication, the most severe complication was selected. In-hospital mortality was defined as any patient with postoperative death at the time of discharge within a 30-day hospital stay. Hospital stay was defined as the time from initial hospitalization postoperatively to discharge. In this study, EURO refers to reoperation inside the peritoneal cavity performed in a patient with GC under general anesthesia and tracheal intubation because severe postoperative complications were caused by the original LAG or OG procedure within a 30-day hospital stay. The EURO rate was defined as the number of patients requiring reoperation divided by all of the patients.

Statistical analysis

Statistical analyses were performed using SPSS version 18.0 (SPSS, Chicago, IL, USA). Categorical data are presented as percentages and were analyzed using the χ2 test or Fisher’s exact test, whereas continuous data are reported as the mean ± standard deviation (SD), and the differences in the means between the groups were analyzed using Student’s t-test. Variables with a p value of less than 0.05 were considered statistically significant. Variables with p < 0.05 in the univariate analysis were subsequently selected for inclusion in a multivariate binary logistic regression model and were presented as hazard ratios (HRs) with 95% confidence intervals (CIs).

Results

Clinicopathological characteristics in all patients and EURO patients

The clinicopathologic characteristics of all patients (n = 4124) are shown in the Supplementary Table. Statistical differences were observed in age (p < 0.001), BMI (p = 0.019), comorbidities (p = 0.014), ASA score (p < 0.001), malnourishment (p < 0.001), tumor size (p < 0.001), history of abdominal surgery (p = 0.007), tumor location (p < 0.001), reconstruction type (p < 0.001), T stage (p < 0.001), N stage (p < 0.001), and TNM stage (p < 0.001) between the LAG and OG groups. There was no statistically significant difference in gender (p = 0.655) or the extent of gastrectomy (p = 0.254) between the two groups.

The clinicopathological characteristics of the patients requiring EURO (n = 52) are shown in Table 1. Among 52 patients requiring EURO, 44 (84.6%) were male, and 8 (15.4%) were female. The mean age was 63.44 ± 10.78 years, and the mean BMI was 22.97 ± 3.97 kg/m2. There were no significant differences in gender, age, BMI, comorbidities, ASA score, malnourishment, tumor size, history of abdominal surgery, tumor location, reconstruction type, extent of gastrectomy, N stage, or TNM stage (all p > 0.05) for patients requiring EURO between the LAG and OG groups, but there was a significant difference between the two groups in T stage (p = 0.007).

Table 1 Clinicopathological characteristics of EURO
Full size table

Postoperative complications and EURO

Among 4124 patients, overall postoperative complications within the 30-day hospitalization period occurred in 374 cases (15.2%), and there was a significant difference in the overall postoperative complication rate between the LAG and OG groups (14.3% vs 16.7%, p = 0.043). Severe postoperative complications within the 30-day hospitalization period were observed in 159 cases (3.9%), and there was no statistical significance in the rate of severe postoperative complications (3.8% vs 4.0%, p = 0.669) between the two groups. Approximately 0.6% of patients (23/4,124) died postoperatively within the 30-day hospitalization period, and mortality was higher in the OG group than in the LAG group (0.2% vs 1.1%, p < 0.001). EURO within the 30-day hospitalization was observed in 52 cases (1.3%), and no significant difference was observed in the EURO rate between the LAG and OG groups (1.1% vs 1.6%, p = 0.157). The EURO rate within 24 h postoperatively was similar between the two groups (0.4% vs 0.2%, p = 0.279); however, the EURO rate after 24 h postoperatively was statistically lower in the LAG group than in the OG group (0.7% vs 1.4%, p = 0.019) (Table 2).

Table 2 Morbidity and mortality after LAG and OG
Full size table

Univariate and multivariate analyses

As shown in Table 3, the univariate analysis of all 4124 cases demonstrated that no significant correlation was found between gender (p = 0.107), comorbidities (p = 0.792), ASA score (p = 0.318), malnourishment (p = 0.424), neoadjuvant treatment (p = 0.637), abdominal surgery (p = 0.410), reconstruction type (p = 0.345), gastrectomy type (p = 0.122), and laparoscopy or not (p = 0.157) and EURO. However, age (p = 0.019), BMI (p = 0.006), and estimated blood loss (p = 0.034) were closely related to the EURO rate. Multivariate logistic two-regression analysis revealed that age > 70 (HR 1.950, p = 0.028), BMI > 25 g/m2 (HR 2.288, p = 0.009), and estimated blood loss > 100 ml (HR 1.867, p = 0.029) were independent risk factors for EURO after gastrectomy for GC.

Table 3 Univariate and multivariate analysis of risk factors for EURO
Full size table

Interval time to EURO and causes of EURO

The mean interval time to EURO after the first operation was 6.9 ± 6.7 days, and there was no significant difference in the interval time to EURO between the LAG and OG groups (5.6 ± 5.5 days vs. 8.3 ± 7.8 days, p = 0.148). The leading causes for EURO in the 52 patients requiring EURO were abdominal bleeding in 23 cases (44.2%), anastomotic bleeding in 7 (13.5%), anastomotic leakage in 6 (11.5%), abdominal infection in 5 (9.6%), intestinal obstruction in 5 (9.6%), incision dehiscence in 3 (5.8%), jejunal perforation in 2 (3.8%), and pancreatic fistula in 1 (1.9%). The main cause of EURO in the LAG group was intra-abdominal bleeding, followed by anastomotic bleeding and anastomotic leakage; however, the main cause of EURO in the OG group was intra-abdominal bleeding, followed by anastomotic leakage and intestinal obstruction.

The proportion of EURO patients with abdominal bleeding was significantly higher in the LAG group than in the OG group (57.1% vs. 29.2%, p = 0.043), and there was no significant difference in the proportion of EURO patients with anastomotic bleeding (17.9% vs. 8.3%, p = 0.551), anastomotic leakage (7.1% vs. 16.7%, p = 0.525), abdominal infection (7.1% vs. 12.5%, p = 0.856), intestinal obstruction (3.6% vs. 16.7%, p = 0.261), incisional dehiscence (3.6% vs. 8.3%, p = 0.891), jejunal perforation (3.6% vs. 4.2%, p = 1.000), or pancreatic fistula (0 vs. 4.2%, p = 0.462) between the two groups (Table 4). Among 23 patients requiring EURO due to intra-abdominal bleeding, transverse mesocolon vascular bleeding and unexplained intra-abdominal bleeding were the leading bleeding sites, followed by pancreatic vascular bleeding and splenic lobar artery bleeding. Transverse mesocolon vascular bleeding and splenic bleeding were the most common bleeding sites among EURO patients in the LAG and OG groups, respectively (Table 5).

Table 4 Interval to EURO and main causes of EURO after LAG and OG
Full size table
Table 5 Location of intra-abdominal bleeding in 23 patients requiring EURO
Full size table

Clinical courses of patients undergoing EURO

There were 6 of 28 (21.4%) EURO patients in the LAG group, but none of the 22 EURO patients in the OG group underwent a laparoscopic procedure (p = 0.025). There was no significant difference in the number of EUROs performed between the two groups (p = 0.590). The morbidity rate in EURO patients was 39.2% in the LAG group and 50.0% in the OG group (p = 0.438), and the mortality rate in EURO patients was 3.6% in the LAG group and 20.8% in the OG group (p = 0.084). However, there was no significant difference in the length of hospital stay among EURO patients between the two groups (26.9 ± 14.3 days vs. 29.5 ± 11.6 days, respectively, p = 0.429) (Table 6).

Table 6 Clinical courses of patients with EURO
Full size table

Treatment and recovery after EURO

Management methods in the 52 patients who underwent EURO included laparoscopic surgery in 6 of 28 patients in the LAG group due to intra-abdominal bleeding in 3 cases, anastomotic bleeding in 2 cases, and intestinal perforation in 1 case. The remaining 46 patients (22 in the LAG group and 24 in the OG group) underwent open surgery. Detailed management strategies performed during the 52 EUROs included exploratory surgery alone, suturing or ligation, and splenectomy for intra-abdominal bleeding, endoscopy-guided suturing and debridement and/or reconstruction for anastomotic bleeding, debridement and/or splenectomy for intra-abdominal infection, enterolysis, enterectomy and/or reconstruction for intestinal obstruction, return and relaxation sutures for incisional dehiscence, debridement and repair for intestinal perforation, and debridement and drainage for a pancreatic fistula. Forty-six of 52 patients who underwent EURO recovered smoothly after EURO, but 6 patients died after EURO. Among them, one patient in the LAG group and 2 patients in the OG group died from sepsis, and 2 patients and 1 patient in the OG group died from disseminated intravascular coagulation (DIC) and multiple organ failure, respectively (Table 7).

Table 7 Treatment and Status of EURO
Full size table

Discussion

Postoperative complications have always been a focal point of concern for surgeons. EURO is generally considered necessary for the emergency management of particular, severe postoperative complications. In the previously published literature, the EURO rate after LAG was reported to range from 1.2 to 2.1% [6, 7, 16], and the EURO rate after OG was reported to range from 1.5 to 2.2% [6, 7, 10]. Kim et al. [16] reported that the EURO rate was closely related to preoperative comorbidities, type of reconstruction, operative time, and surgeon’s experience. In a study by Oh [21] of EURO after OG, the results revealed that there was an older mean age and more males in the EURO group than in the non-EURO group. Two large prospective studies that compared the short-term outcomes between laparoscopic and open surgery in gastrectomy patients reported the detailed results of morbidity and mortality but only limited information on EURO itself [6, 7]. However, in contrast to Kim’s [6] and Hu’s [7] studies, in the current study, we focused only on differences in perioperative consequences, treatment strategies and high-risk factors for EURO by comparing patients who underwent LAG or OG to provide surgeons with certain evidence for the prevention and treatment of EURO. The EURO rate was 1.1% in the LAG group and 1.6% in the OG group, which is consistent with the results of previous studies [6, 7, 10, 16]. The type of procedure (laparoscopic or open) was not identified as a risk factor for EURO after gastrectomy for GC; however, age > 70, BMI > 25 kg/m2, and estimated blood loss > 100 ml were independent risk factors for EURO. Therefore, no significant correlation was found between laparoscopic surgery and the EURO rate after LAG. Surgeons should pay particular attention to GC patients with advanced age, high BMI, and a large volume of intraoperative blood loss and improve perioperative management to prevent or reduce the occurrence of EURO after gastrectomy. Interestingly, further stratified analyses showed that the EURO rate within 24 h postoperatively was similar between the two groups, yet the EURO rate after 24 h postoperatively was statistically lower in the LAG group than in the OG group.

The main reasons for EURO after gastrectomy were different in previously published studies. In Kim’s LAG study [16], abdominal bleeding, duodenal stump, and intestinal obstruction were the primary causes of EURO. However, Sah et al. [10] revealed that the leading causes for EURO after OG were abdominal bleeding and anastomotic leakage. In addition, Yi et al. [22] showed that the splenic hilar vessels, transverse mesocolonic vessels, greater omental vessels, and the spleen itself were common bleeding sites leading to EURO after OG. In the current study, the main cause of EURO was intra-abdominal bleeding, followed by anastomotic bleeding, anastomotic leakage in the LAG group, and intra-abdominal bleeding, followed by anastomotic leakage and intestinal obstruction in the OG group. The transverse mesocolonic vessels and spleen were the most common bleeding sites requiring EURO in the LAG and OG groups, respectively.

In this study, interestingly, we also found that the EURO rate within 24 h postoperatively was similar between the two groups, yet the EURO rate after 24 h postoperatively was statistically lower in the LAG group than in the OG group. Moreover, the proportion of patients with abdominal bleeding requiring EURO was markedly higher in the LAG group than in the OG group. Both in the LAG and OG groups, the main causes for EURO within 24 h after the initial surgery were intra-abdominal bleeding and anastomotic bleeding. On the one hand, we believe that improper application of energy devices [23, 24] intraoperatively, including the insufficient or excessive coagulation of blood vessels during LAG, probably causes bleeding due to blood clots breaking off the cut end of the vessels within 24 h after LAG, which may partly result in a higher EURO rate in the LAG group than in the OG group because of postoperative intra-abdominal bleeding. On the other hand, we purport that surgeons fully believe in and rely too much on all types of anastomotic instruments during LAG, frequently ignoring to inspect the anastomosis after reconstruction of the digestive tract, which may in part lead to the occurrence of EURO because of postoperative anastomotic bleeding within 24 h after LAG. Therefore, surgeons should attach great importance to the correct use of energy devices and anastomotic instruments, and reliable hemostasis and careful inspection of divided vessels and reconstructed anastomoses during surgery. This can help to reduce the rate of intra-abdominal bleeding or anastomotic bleeding within 24 h after surgery, which will further decrease the rate of EURO after LAG.

Laparoscopic surgery is safe and feasible for the treatment of patients with a history of abdominal surgery [25,26,27]. However, there are no relevant reports on whether laparoscopic procedures can be performed to treat emergency patients who require EURO after LAG. In the current study, laparoscopic surgery was successfully performed in 21.4% of patients in the LAG group requiring EURO. The management process of reoperation was similar between the LAG and OG groups and included suturing or ligation for hemostasis to control bleeding, splenectomy for hemorrhage, debridement plus irrigation-drainage or reanastomosis plus irrigation-drainage for anastomotic leakage and so on. For experienced surgeons, therefore, laparoscopy procedures could still be applied as an alternative to treat some patients requiring EURO after the initial LAG procedure for GC. Previous studies [10] have reported that the mortality rate after EURO following OG for GC was as high as 11.1%. In the current study, the mortality rate after EURO in the LAG group was 3.6%, which was lower than the mortality rate of 20.8% in the OG group, with marginal significance. The lower postoperative mortality rate after EURO in the LAG group may be mainly related to the laparoscopic procedure itself used for the initial operation, and EURO is associated with minimal surgical trauma and reduced stress response among the patients [28, 29].

For the first time, this study provides some valuable insights that can aid surgeons when planning to perform LAG or OG for GC. However, certain limitations also exist in this study. First, as a single-center retrospective study, the small sample size of patients who underwent EURO after gastrectomy for GC is one of the limitations of this study. Second, the average BMI of the patients in this study was relatively low compared to patients in other Western studies. Third, there were some differences in the clinicopathologic characteristics of the patients who underwent open vs. laparoscopic surgery in this study, partly because laparoscopy is performed more commonly than the open procedure, which may have increased the selection bias of patients to some extent. Fourth, patients were eligible for inclusion in the OG group and the LAG group in this study from Jan. 2005 to Dec. 2014 and from Jul. 2007 to Dec. 2014, respectively; thus, there was an obvious time difference during which patients were included in the two groups. However, Jul. 2007 was the cut-off date, and the patients who underwent OG were included in the early group, in which there were 658 cases from Jan. 2005 to Jul. 2007, and the late group, in which there were 858 cases from Aug. 2007 to Dec. 2014. The EURO rates between the early and late groups were 1.1% and 2.0%, respectively, without a significant difference (p = 0.156). Thus, prospective, multicenter, large-sample studies should be performed in the future to help confirm all these observations.

In conclusion, no significant correlation between laparoscopic surgery and the EURO rate was found in patients undergoing gastrectomy for GC. Laparoscopic surgery was associated with a lower EURO rate after 24 h postoperatively and a higher proportion of patients undergoing EURO for intra-abdominal bleeding than open surgery. Effective and accurate intraoperative hemostasis for intra-abdominal vessels and anastomotic sites will help to further reduce the incidence of EURO following LAG within 24 h postoperatively.