We retrospectively reviewed the cases of 303 patients who underwent LTG for middle- or upper-third gastric carcinoma between 2005 and 2016 at Chonnam National University in South Korea. Patients who had previously undergone major abdominal operations (hepatobiliary or colorectal surgery) or had undergone combined major organ resection (spleen, kidney, or adrenal gland) were excluded from the study to rule out the effects of these procedures on the postoperative outcomes. All patients underwent upfront surgery without preoperative chemotherapy. The surgeries were conducted by 4 surgeons who performed about 150 to 700 operations by the time the study was initiated. The decision to perform LTG was made based on preoperative staging. LTG was originally indicated for patients with cT1N0 gastric carcinoma; however, as surgeons gain experience with laparoscopic surgery, they also performed LTG on patients with more advanced gastric cancers, such as cT2-3N0-2, during the latter stages of the study period. This study was approval by the Institutional Review Board at Chonnam National University Hwasun Hospital in South Korea (CNUHH-2017-153) was performed in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. For this type study, formal consent is not required.

The patients' demographic data, clinicopathological characteristics, and information on their surgical outcomes were retrospectively collected from our institution's database collection. Tumor characteristics and pathological stages were based on the guidelines of the Japanese classification of gastric carcinoma [12] and the 7th edition of the Union for International Cancer Control tumor, node, metastasis (TNM) classification [13], respectively. Postoperative complications were defined as those that occurred within 30 days after the surgery and were classified as local or systemic based on whether or not they occurred within the operating field. The severity of postoperative complications was graded based on the Clavien-Dindo classification [14].

Patients with a history of medical treatment for any disease were considered to have a comorbidity. Information regarding such comorbidities was obtained from medical history records or during the preoperative workup; the latter included complete blood count, coagulation, biochemical, and pulmonary function tests, electrocardiography, and chest radiography. If abnormal findings were observed in the preliminary workup, further appropriate tests were performed to diagnose any underlying diseases. We investigated the most common comorbidities including hypertension, diabetes mellitus, chronic viral hepatitis, liver cirrhosis, and pulmonary, ischemic heart, and cerebrovascular diseases; other less common types of comorbidities were recorded, as appropriate. Patients with abnormal preoperative pulmonary function tests that suggested the presence of chronic obstructive pulmonary disease (COPD), i.e., a forced expiratory volume in 1 second/forced vital capacity ratio <0.7, or those with a history of pulmonary disease (COPD, asthma, or pulmonary tuberculosis) were considered to have pulmonary disease.

About 5–6 abdominal ports were used for LTG. The surgeon stood on the right side of the patient. After a pneumoperitoneum of 12 to 14 mmHg was achieved, gastric dissection was initiated by dividing the greater omentum and moving towards the left gastroepiploic area. All gastric and lymph node dissections were performed with the use of laparoscopic energy devices such as the Harmonic Scalpel™ (Ethicon Endo-Surgery, Cincinnati, OH, USA) or Ligasure™ (Valleylab, Boulder, CO, USA). The extent of lymph node dissection followed the Japanese treatment guidelines for gastric carcinoma; D2 lymph node dissection was indicated for patients with clinical stage ≥cT2 or cN+ disease [15]. The lymph nodes in the splenic hilum are usually dissected to preserve the spleen (i.e., the spleen-preserving technique).

Esophagojejunal anastomosis was performed via either the extracorporeal or intracorporeal method. For the extracorporeal method, esophagojejunal reconstruction was performed using a 25-mm circular stapler through a 6–7 cm mini-laparotomy on the epigastrium, which is similar to the conventional open technique. On the other hand, for the intracorporeal method, side-to-side esophagojejunostomy (overlap method) was performed using a 45-mm linear stapler after resecting the abdominal esophagus with a linear stapler. A common hole at the esophagojejunostomy site was closed using the hand-sewn method or a linear stapler.

A standardized care protocol was utilized in the management of patients. Early postoperative oral nutrition has been implemented since 2009, wherein an oral diet was initiated on postoperative day 1 or 2. Intravenous fluid restriction (20–25 mL/kg/day for 3 to 4 postoperative days) was performed in patients receiving early oral nutrition. Preoperative mechanical bowel preparation or nasogastric tube insertion was not routinely performed. Additionally, abdominal drainage was inserted in selected cases based on the surgeons' decision. Patients were discharged from the hospital on postoperative day 6 or 7 based on the predefined discharge criteria.

Data are expressed as a mean±standard deviation or number (%). Continuous variables were compared using a t-test while categorical variables were compared using the Fisher's exact or χ² test as appropriate. The binary logistic regression model was used for univariate and multivariate analyses of the risk factors for postoperative complications. In the logistic regression model, continuous data were analyzed after being categorized based on their median values. Variables that were statistically significant in the univariate model were subjected to multivariate analysis. All statistical analyses were performed using the SPSS version 23.0 software (IBM Corp., Armonk, NY, USA). For all analyses, 2-sided P-values <0.05 were considered statistically significant.

This study included 212 men and 91 women with a mean age of 62.1±11.0 years and mean body mass index of 23.6±3.2 kg/m² (Table 1). Among them, 260 (85.8%) and 43 (14.2%) patients underwent extracorporeal and intracorporeal anastomoses, respectively. The mean length of hospital stay was 9.9±8.4 days. Based on the final pathological examination, 254 (83.8%), 36 (11.9%), and 13 (4.3%) patients had stages I, II, and III disease, respectively.

A total of 189 patients (62.4%) had comorbidities (Table 1); 89 patients (29.4%) had ≥2 comorbidities. The most common comorbidity was hypertension (37.0%) followed by diabetes mellitus (17.8%), chronic viral hepatitis (3.0%), liver cirrhosis (2.6%), and pulmonary (27.1%), ischemic heart (3.3%), and cerebrovascular diseases (2.3%). Among the 82 patients with pulmonary disease, 65 of them showed abnormal preoperative pulmonary function tests indicative of COPD while 16 patients had a history of pulmonary tuberculosis. Active treatment for COPD or asthma was conducted in 21 patients.

Table 2 shows the details of the postoperative complications. The overall morbidity and mortality rates were 20.1% and 1.0%, respectively. The ≥grade 3 complications occurred in 20 patients (6.6%). Three patients died including 1 each with anastomosis leakage, abdominal infection, and luminal bleeding. Luminal bleeding (n=16) was the most common local complication, followed by anastomosis leakage (n=11), paralytic ileus (n=9), and abdominal infection (n=8). Pulmonary complication was the most common systemic complication.

A significant difference was noted in the incidence of the local (P=0.021), overall (P=0.006), and ≥grade 3 complications (P=0.025) and the number of comorbidities increased (Fig. 1). However, no significant difference was noted in the incidence of systemic complications. The postoperative complication rates in patients with 7 major types of comorbidities are shown in Fig. 2. Patients with pulmonary disease showed significantly higher complication rates than those without comorbidities (32.9% vs. 14.9%, respectively, P=0.003). On the other hand, no significant differences were observed between patients with other comorbidities and the incidence of postoperative complications.

To further investigate the impact of each comorbidity on postoperative complications, we performed univariate and multivariate analyses of the risk factors for postoperative complications (Table 3). Results of the univariate analysis showed that old age, male sex, the number of comorbidities, presence of pulmonary disease, and intraoperative blood loss volume were significantly associated with postoperative complications. Multivariate analysis of these factors revealed that pulmonary disease was an independent risk factor for postoperative complications (odds ratio, 2.14; 95% confidence interval, 1.03–4.64) along with old age and intraoperative bleeding.

To identify the type of complications which increased because of pulmonary disease, we compared the incidence of each complication in patients with pulmonary disease to that in patients without comorbidities (Table 4). Luminal bleeding was found to be significantly more common in patients with pulmonary disease than in those without any comorbidities (12.2% vs. 3.5%, P=0.020). Most luminal bleeding events in patients with pulmonary disease occurred during esophagojejunal anastomosis (n=9), and all patients were successfully treated with conservative treatment or endoscopic intervention. There were no significant differences in the rates of other complications between the 2 groups.