Risk factors of postoperative complications and their effect on survival after laparoscopic gastrectomy for gastric cancer

Abstract Background The association between postoperative complications and long‐term survival after laparoscopic gastrectomy (LG) for gastric cancer (GC) remains uncertain. This study aimed to determine the incidence and risk factors of postoperative complications and evaluate their impact on survival outcomes in patients undergoing LG. Methods A retrospective study was conducted on 621 patients who underwent LG for gastric adenocarcinoma between March 2015 and December 2021. Postoperative complications were classified according to the Clavien–Dindo classification, with major complications defined as Grade III or higher. Logistic regression models with stepwise backward procedure were used to identify risk factors for complications. To assess the impact of postoperative complications on survival, uni‐ and multi‐variable Cox proportional hazard models were used for overall survival (OS) and disease‐free survival (DFS). Results Overall rate of postoperative complications was 17.6% (109 patients); 33 patients (5.3%) had major complications. Independent risk factors for major complications were Charlson comorbidities index (OR [95% CI], 1.87 [1.09–3.12], p‐value = 0.018 for each one score increase), and type of anastomosis (OR [95% CI], 0.28 [0.09–0.91], p‐value = 0.029 when comparing Billroth II with Billroth I). Multivariable analysis identified major complications as an independent prognostic factor to reduce OS (HR [95% CI], 2.32 [1.02–5.30], p‐value = 0.045) and DFS (HR [95% CI], 2.63 [1.37–5.06], p‐value = 0.004). Other prognostic factors for decreased survival outcomes were tumor size, presence of invasive lymph nodes, and T4a stage. Conclusions Major complications rate of LG for GC was approximately 5.3%. Charlson comorbidities index and type of anastomosis were identified as risk factors for major postoperative complications. Major complications were demonstrated to pose adverse impact on survival outcomes.


| INTRODUC TI ON
Gastric cancer (GC) is a prevalent malignancy and a leading cause of cancer-related mortality, representing a significant global health burden across the world. 1 Despite notable advancements in multimodality treatment and targeted therapy, gastrectomy remains the primary treatment option for this disease.Laparoscopic radical gastrectomy is commonly employed for early GC and is increasingly being utilized for advanced GC.This procedure has become the standard approach for proficient surgeons.[4][5][6] Radical gastrectomy plays a crucial role in improving survival outcomes of GC patients.However, this surgical technique is intricate and challenging, leading to a high incidence of postoperative complications.Achieving proper lymph node dissection and R0 resection are essential components of radical gastrectomy.][5][7][8][9] Major postoperative complications, such as abdominal abscess, bleeding, anastomotic leak, pneumonia, can contribute to increased treatment costs, prolonged hospitalization, and an elevated risk of recurrence and worse long-term survival.[12][13][14][15][16] While surgical techniques and surgeon expertise have improved, managing postoperative complications remains a persistent challenge.The Clavien-Dindo classification system serves as a universally accepted standard for assessing postoperative complications. 2,8,9,14The rate and severity of postoperative morbidity can be utilized as a metric to evaluate the quality of surgery.Identifying preoperative risk factors is crucial for mitigating postoperative complications. 7,17,18[21][22] Nonetheless, the association between postoperative complications and long-term survival after LG for GC remains a subject of debate.
In this study, we aimed to examine the incidence of postoperative complications following LG and lymphadenectomy using the Clavien-Dindo classification system.Additionally, we sought to analyze the risk factors associated with these complications and the impact of complications on survival outcomes in patients with GC within a single institution.We started performing laparoscopic proximal gastrectomy with double flap reconstruction in 2018 and did only four cases for early GC.Therefore, we excluded this procedure in this study to avoid bias.
All patients included in the study were confirmed to have GC through preoperative endoscopic biopsy and histological analysis.
Additionally, abdominal computed tomography (CT) scan was performed to assess the stage and location of the tumor.The indication for LG was cT1b-T4a, N0-3, M0, and cT1aN0M0 with a failed endoscopic submucosal resection.

| Surgical techniques
Surgical techniques applied in this study followed the principles of laparoscopic distal gastrectomy (LDG) and laparoscopic total gastrectomy (LTG), in accordance with the Japanese guidelines for the extent of gastrectomy and D2 lymphadenectomy. 23During the procedures, total omentectomy was performed in all cases.The anastomosis was performed using intracorporeal techniques with a linear stapler.In LDG, the anastomosis methods included Billroth-I, Billroth-II, and Roux-en-Y, while in LTG, Roux-en-Y was utilized for all cases.Mobilization and dissection were conducted using an ultrasonic scalpel.Routine peritoneal lavage cytology was not performed.or TS-ONE®.During the first 2 years after surgery, patients were followed up every 3 months, followed by biannual visits for the subsequent 3 years, and then annual visits.During each follow-up visit, patients underwent physical examinations, laboratory blood tests, and abdominal ultrasonography.CT scans were conducted every 6 months during the first 3 years and annually thereafter.Endoscopy was performed annually.In cases where patients exhibited suspected symptoms or signs of recurrence or metastasis, CT scans and/or endoscopy were performed to evaluate and confirm the presence of such conditions, regardless of the scheduled follow-up appointments.The length of follow-up was defined as the duration from the date of surgery to the final follow-up date in December 2022 or the date of death.Cancer recurrence was diagnosed based on radiologic or histological evidence of disease.

| Outcomes
Postoperative complications were evaluated within 30 days after LG using the Clavien-Dindo classification system, which consists of five main grades. 25 Postoperative mortality was defined as any death occurring during the hospital stay following surgery.Pneumonia was confirmed by clinical pyrexia and infiltration on the chest X-ray.
Long-term oncological outcomes were assessed in terms of OS and DFS during the follow-up period.OS represents the duration from surgery to death from all causes, while DFS measures the length of time from surgery to the detection of recurrence or metastasis, or death from any cause.If no events (recurrence, metastasis, or death) occurred, OS and DFS were censored at the last follow-up time.

| Statistical analysis
To investigate risk factors of postoperative complication, we conducted separate analyses for overall complication and major complication, both of which were binary variables (Yes/No).Minor complications were included in the "Yes" group for the overall complications but were included in the "No" group for the major complication analysis.Univariable analysis was performed using two-sample t-test for normally distributed numeric variables, the Wilcoxon rank-sum test for non-normally distributed numeric variables, and Fisher's exact test for categorical variables.Multivariable analysis was conducted using logistic regression models with a stepwise backward procedure to identify independent risk factors of postoperative complications.
Survival outcomes were presented using Kaplan-Meier plots stratified by the level of complications (no complications, minor complications, and major complications).To assess the impact of postoperative complications on survival, uni-and multivariable Cox proportional hazard models were used for OS and DFS analyses, including complication with three levels and other potential confounding factors.
All analyses were performed using R statistical software version 4.1.0.Two-sided tests were used for all analyses, and a p-value less than 0.05 was considered statistically significant.The results from the models were reported as odds ratios (ORs) for logistic regression models and hazard ratios (HRs) for Cox models, along with their corresponding 95% confidence intervals (CIs).

| RE SULTS
Between March 2015 and December 2021, a total of 712 patients with gastric adenocarcinoma underwent LG in our center.After excluding 91 patients, 621 patients were included in this study (Figure 1).The distribution of patients by period was as follows: 2015-2017-177 patients; 2018-2019-226 patients; and 2020-2021-218 patients.

| Clinical characteristics
Table 1 presented the patients' clinical characteristics.The mean age was 59.5 ± 11.9 years, and 72.2% were of normal weight.The most common comorbidity was hypertension (24%), followed by diabetes (10.6%) and ischemic heart disease (5.8%).Forty-four patients (7.1%) had a history of prior abdominal surgery.Most patients (87.6%) were classified as ASA-PS classification I or II.
Regarding tumor characteristics, the most common location of the tumors was the lower third of the stomach (68%), followed by the middle third.The mean tumor size was 3.6 cm.Most patients (61.8%) were diagnosed at the locally advanced stage with serosa invasion (T4a).Preoperative chemotherapy was administered to 2.9% of the patients.LDG was performed in 530 patients (85.3%), while LTG was performed in 91 patients (14.7%).The Roux-en-Y reconstruction method was utilized in all 91 LTG cases and in 139 LDG cases.
The most used reconstruction method for LDG was the Billroth-II.
Among the patients, the length of postoperative stay was relatively longer in the major group (average: 12 [10;19]

| Risk factors for postoperative complications
In the univariable analyses, factors including preoperative anemia,  2).
Regarding the major postoperative complications, the results of the univariate analysis revealed that operating time, type of surgery, and type of anastomosis were associated with a higher rate of major complications.Furthermore, the multivariable analysis iden- Billroth II with Billroth I) as independent risk factors for the development of major complications (Table 3).

| Postoperative complications and survivals
During a median follow-up period of 30 months, disease progression occurred in 104 patients, and 99 patients died.The survival rate of both the minor and major complication groups appeared to be lower compared to the group with no complications (Figure 2).The 3-year OS rates (excluded patients with preoperative chemotherapy) for the no complication, minor complication, and major complication groups were 86%, 77%, and 79%, respectively.The 3-year DFS rates were 78%, 67%, and 61% for the respective groups (Table S2).The most common type of recurrence was peritoneal (62 cases, 59.6%), followed by hematogenous (15 cases, 14.4%) (Table S3).
Regarding Regarding the other prognostic factors for decreased survival outcomes, the results of the multivariable analysis demonstrated that tumor size, presence of invasive lymph nodes, and T4a stage were also independent risk factors of decreased OS and DFS (Table 4, Table 5).

| DISCUSS ION
Recently, LG became increasingly common for the surgical treatment of AGC due to its reported benefits, including reduced blood loss, less postoperative pain, and early return to normal bowel function.Our study demonstrated that LG for GC was safe and feasible.We observed an overall postoperative complications rate of 17.6%, a major complication rate of 5.3%, and a 30-day postoperative mortality rate of 0.5%, which aligned with previous studies.Note: Statistical summary is mean ± standard deviation, median (25th; 75th percentiles), or n (%).
Abbreviations: ASA-PS, The American Society of Anesthesiologists physical status; BMI, body mass index; CI, confidence interval; HR, hazard ratio; Ref, reference.

TA B L E 3 (Continued)
Anastomotic leakage was a significant concern in gastrointestinal surgery and could lead to other complications such as bleeding, abdominal infection, peritonitis, and septic shock, which carried a risk of mortality.In our study, we observed an anastomotic and duodenal stump leak rate of 1.6%, in which two cases unfortunately resulted in septic shock and death.Our leakage rate was similar to recent TA B L E 4 Uni-and multivariable analyses of risk factors of overall survival (excluded patients with preoperative chemotherapy).studies 2,5,8,9,26 and lower compared to earlier studies 27,28 possibly due to the advancements in surgical techniques and instrument technology.
Pancreatic fistula was another common postoperative complication associated with significant inflammation and prolonged hospital stays.
In our study, most pancreatic fistula cases were minor complications, with only one patient experiencing a major complication.
TA B L E 5 Uni-and multivariable analyses of risk factors of disease-free survival (excluded patients with preoperative chemotherapy).LG.These findings were heterogenous to prior studies. 2,7,9,21,29veral studies indicated that postoperative complications were associated with operative time and blood loss (intraoperative bleeding). 2,7,9,21,29Contrarily, our study found these factors to have a non-adverse impact.Advances in anesthesia techniques and enhancements in perioperative management have contributed to a reduction in complication rates among patients with unfavorable intraoperative conditions.It is worth mentioning that all comorbidities were effectively managed prior to surgery, and we implemented the ERAS program as a standard practice.Despite these measures, the Charlson comorbidities index still had an impact on the incidence of both overall and major postoperative complications, which is consistent with previous studies. 2,9,21,291][32] Consistent with these findings, our study demonstrated that major complications had an impact on long-term OS and DFS.It is crucial to closely monitor patients for potential major complications and take prompt action to prevent further morbidity and mortality.The negative association between major postoperative complications and poor survival after LG for GC can be attributed to the postoperative inflammatory response, which contributes to the host immunosuppression.This immunosuppression compromises cell-mediated immunity, particularly affecting natural killer cells and cytotoxic T lymphocytes, thereby promoting the proliferation and metastasis of residual tumor cells. 33Additionally, major postoperative complications may result in lengthening postoperative hospital stay, delays or omission of adjuvant chemotherapy, further exacerbating the adverse impact on survival outcomes following LG for GC.In our study, the length of postoperative hospital stay was longer in the major complication group (average: 12 [10;19] days) compared to the no complication group (average: 7 [7;8] days).Moreover, the rate of patients receiving complete adjuvant chemotherapy was lower in the major complication group (55.6%) compared to the no complication group (68.4%).Previous research has showed that patients who received complete adjuvant chemotherapy had better survival outcomes compared to those who underwent gastrectomy alone. 34jor postoperative complications could significantly impact long-term survival after laparosopic gastrectomy.However, other factors have also contributed to this scenario.Consistent with the findings of numerous previous studies, 6,8,9,13,15,16,19,27,30,31 our results also demonstrated that tumor size, the presence of invasive lymph nodes, and T4a stage were as important factors in predicting survival of both OS and DFS.In general, patients presenting with these tended to have lower survival rates and higher rates of postoperative complications compared to those with less advanced tumors.
In our study, the median length of follow-up was more than 2.5 years.Therefore, estimates for survivals over 3 years are not certain.However, the Kaplan-Meier (Figure 2) showed the proportional hazard rates of the three groups (no complication, minor complication, and major complication).With longer follow-up, we believe that the difference between groups still holds with statistical significance.Additionally, longer follow-up is required to confirm this finding.
The study has several limitations that should be acknowledged.
Firstly, this study was conducted at a single center, which may limit the generalizability of the findings to other settings or populations.
Secondly, the relatively small number of specific complications ob- In conclusion, the study provided evidence that LG was a safe and feasible surgical approach for GC.The incidence of major complications was found to be 5.3%.The Charlson comorbidities index was identified as an independent risk factor for both overall and major postoperative complications.Type of anastomosis was another risk factor for major complications.Importantly, major complications posed negative impact on survival outcomes.Therefore, efforts should be made to minimize the occurrence of complications and ensure timely and appropriate multimodal therapies for patients with complications to improve survival of GC patients.

2 | PATIENTS AND ME THODS 2 . 1 |
Study design and populationThis retrospective study was conducted at the Gastro-Intestinal Surgery Department of the University Medical Center Ho Chi Minh City, a referral hospital in Southern Vietnam.We included all patients who underwent LG for GC between March 2015 and December 2021.The inclusion criteria for the study were as follows: (i) histologically confirmed adenocarcinoma of the stomach; (ii) surgical staging of sT1-T4a, N0-3, M0 based on the 7th edition of the American Joint Committee on Cancer/Union Internationale Contre le Cancer (AJCC/UICC) staging system.Exclusion criteria encompassed the following: (i) proximal gastrectomy; (ii) intraoperatively detected bulky lymph nodes; (iii) inadequate lymphadenectomy (D0, D1); (iv) presence of macroscopic residual tumor (R2); (v) concurrent cancer or a history of previous other cancers; (vi) previous gastrectomy, (vii) palliative gastrectomy; (viii) complications such as bleeding or perforation requiring emergency gastrectomy, (ix) lost to follow-up within 30 days after surgery.

2 . 3 |
All operations were performed by three experienced surgeons, each with a track record of over 200 standard open and laparoscopic gastrectomies for GC.We have performed laparoscopic gastrectomy K E Y W O R D S gastric cancer, laparoscopic distal gastrectomy, laparoscopic gastrectomy, laparoscopic total gastrectomy, postoperative complications for early GC since 2008 and for locally advanced GC since 2012.The technique of LG and lymph node dissection had been standardized for advanced GC since 2013.All surgeons in this study achieved an adequate learning curve since 2013.Prophylactic antibiotics were administered as a routine practice.Abdominal drains were routinely inserted and removed within 2-3 days after surgery.The use of nasogastric tube was not standard practice.Since 2017, the enhanced recovery after surgery (ERAS) program has been implemented for all LDG cases.Patient evaluation and follow-up Patient characteristics including gender, age, body mass index (BMI), previous abdominal surgery, American Society of Anesthesiologists physical status (ASA-PS) classification, presence of preoperative anemia, tumor location, tumor size, receipt of preoperative chemotherapy, operating time, blood loss, type of surgery, type of anastomosis, level of lymphadenectomy, differentiation classification, presence of invasive lymph nodes, tumor stage, and Charlson comorbidities index 24 were collected to assess the risk factors for postoperative complications.The follow-up schedule adhered to the Japanese guidelines for GC management.Patients with pathological stage II or higher received adjuvant chemotherapy with either capecitabine and oxaliplatin (XELOX) or fluorouracil, leucovorin, and oxaliplatin (FOLFOX) Grade I complications do not require any pharmacological treatment or intervention.Grade II complications necessitate pharmacological therapy.Grade III complications require surgical, endoscopic, or radiological intervention under regional/local anesthesia (III-a) or general anesthesia (III-b).Grade IV complications indicate life-threatening conditions that require immediate care/intensive care unit management with either single organ dysfunction (IV-a) or multiple organ dysfunctions (IV-b).Grade V is assigned in the event of death.Major complications were defined as Grade III or higher.Diagnosis of anastomotic leak or duodenal stump leak was based on clinical symptoms and signs of peritonitis, along with radiologic examinations demonstrating contrast leakage into abdominal cavity or confirmed during reoperation.Abdominal abscess was confirmed by the presence of a collection of pus, either through percutaneous drainage or during reoperation.Bleeding complications were defined by the requirement of a blood transfusion (two units or more) and clinical symptoms indicating intra-abdominal or gastrointestinal hemorrhage.Pancreatic fistula was diagnosed when any measurable volume of fluid was observed in the drain after the third postoperative day with an amylase content exceeding three times the serum level.Wound infection was diagnosed when purulent exudate was found in the wound, accompanied by positive bacterial culture.

TA B L E 1
Clinical characteristics.

tified
Charlson comorbidities index (OR [95% CI], 1.87 [1.09-3.12],p-value = 0.018 for each one score increase), and type of anastomosis (OR [95% CI], 0.28 [0.09-0.91],p-value = 0.029 when comparing In our study, anastomosis related complications in Billroth I group was significantly higher than in Billroth II group.Notably, anastomotic tension due to the short remnant duodenum and the stomach might be a primary contributing reason.To mitigate this complication, Billroth I anastomosis should be performed in patients with the length of the remnant duodenum at least 2 cm.Additionally, suturing the intersection of the stapler lines is deemed necessary.Surgeons should be noted about the technique and the choice of reconstruction after distal gastrectomy to avoid anastomosis-related complications.Comprehending the pertinent risk factors is crucial for reducing the occurrence of postoperative complications.Our study identified several independent risk factors for postoperative complications.We found that BMI, Charlson comorbidities index, tumor location, tumor size, and advanced T stage were all independent risk factors for overall complications.Additionally, Charlson comorbidity index and type of anastomosis were identified as independent risk factors for major complications following served in the study limited the ability to analyze the risk factors and the impact of each complication on survival outcomes.Further studies with larger sample sizes could provide more detailed insights into these relationships.Finally, the lack of information regarding the starting time of adjuvant chemotherapy limited the ability to fully analyze the association between postoperative complications and delayed initiation of adjuvant chemotherapy, which is known to affect survival outcomes.

analysis Multivariable analysis No (N = 512) Yes (N = 109) p-value OR 95% CI p-value
(Continues) bleeding (33.3%), and abdominal abscess (27.3%) (TableS1).Three patients (0.5%) died during their hospital stay, with two cases attributed to severe septic shock caused by anastomotic leak and one case due to severe bleeding.For anastomosis-related complications after laparoscopic distal gastrectomy, six patients experienced major complications related to Billroth I anastomosis, including intra-abdominal abscess (three patients), anastomotic stenosis (two patients), and major TA B L E 2 (Continued) | 587 LONG et al.TA B L E 3 Uni-and multivariable analysis of risk factors for major complications.
Abbreviations: ASA-PS, The American Society of Anesthesiologists physical status; BMI, body mass index; CI, confidence interval; HR, hazard ratio.