Laparoscopic gastrectomy is becoming a standard procedure for treating early gastric cancer. As the incidence of early gastric cancer increases, laparoscopic gastrectomy has been more widely used, particularly in Asian countries, which have a high incidence of gastric cancer [12].

The safety and efficacy of laparoscopy-assisted distal gastrectomy (LADG) to treat early gastric cancer have been well demonstrated in many clinical studies [345]. An increasing number of LADG procedures have been performed to treat early gastric cancer, with many advantages, including better cosmetic effects, less pain, earlier recovery, shorter hospital stay, and better quality of life than conventional open distal gastrectomy [678].

Totally laparoscopic distal gastrectomy (TLDG) was first introduced in 1992 and demonstrated intracorporeal Billroth II anastomosis using laparoscopic linear staples [9]. In 2002, Kanaya et al. [10] introduced the delta-shaped anastomosis in a TLDG, which is an intracorporeal Billroth I anastomosis using linear staples. TLDG has been used increasingly to treat early gastric cancers with advances in laparoscopic techniques. However, LADG is more preferable than TLDG in many centers, except a few large medical centers, because of difficulties performing an intracorporeal anastomosis and limited experience.

Lymph node dissection could be performed laparoscopically during LADG, but a small laparotomy is necessary for the resection of the stomach and performing the anastomosis. However, TLDG appears to be less invasive than that of LADG because all operative procedures, including gastric resection and reconstruction are performed intracorporeally. Several studies have reported the advantages of TLDG over LADG, such as a smaller incision and less invasiveness [111213], but no controlled randomized study has directly compared TLDG with LADG.

We conducted this study introduce our early experience of TLDG in a single institution and to compare the early surgical outcomes between patients who underwent TLDG and LADG for treatment of early gastric cancer.

Laparoscopic gastrectomy is widely used as a safe and effective procedure to treat early gastric cancer [1415]. Although Goh et al. [9] first introduced TLDG with Billroth II anastomosis in 1992, TLDG received more interest after introduction of the delta-shaped Billroth I anastomosis [10]. The advantage of Billroth I anastomosis is that the alimentary tract structure remains physiologically intact after reconstruction. Thus, we performed the delta-shaped Billroth I anastomosis as the main method for reconstruction during TLDG (31 of 33 cases in the TLDG group). TLDG with the Billroth II anastomosis was performed only when the remnant stomach and duodenum was too far to perform gastroduodenostomy.

There are some limitations in directly comparing the surgical outcomes of LADG and TLDG. For example, the periods that LADG and TLDG were performed were different because LADG was rarely performed after the start of TLDG in many centers, and perioperative clinicopathological status of patients who underwent LADG and TLDG were different because of the small number of cases [1617]. To reduce selection bias according to the discrepancy of operative periods, we matched age, sex, BMI, and ASA score using the PSM method. A case-control analysis was performed using PSM, which is a useful statistical method to reduce bias caused by imbalanced covariates in an observational study. We have found no significant differences in the clinicopathological backgrounds between the LADG and TLDG groups after PSM.

Several studies have compared LADG and TLDG. Previous studies have reported that there was shorter hospital stay in TLDG than that in LADG [1112]. Song et al. [13] reported that bowel recovery time was shorter in patients who underwent TLDG. Kim et al. [18] reported no difference in postoperative clinical course but that postoperative pain was significantly less in patients who received TLDG. In our study, hospital stay in the TLDG group was significantly shorter than that in the LADG group, whereas time to first flatus, start of water, and start of a soft meal were not different between the groups. Although time to flatus and dietary intake could reflect bowel recovery after surgery, numerous factors can affect these results. Therefore, hospital stay is a more reliable variable to reflect early patient recovery. The postoperative complication rate was not different between the groups in this study. However, further study is needed to demonstrate the safety of TLDG compared with that of LADG.

In this study, estimated blood loss in the TLDG group was significantly less than that in the LADG group. A similar result has been reported in many studies [11121319]. It is difficult to explain the reason for less blood loss during TLDG, because the procedure for lymph node dissection was not different between groups. However, advance of surgeon's skill might be a factor that affected this result, because TLDG was performed in later period than LADG in this study. To clarify the advantage of TLDG in terms of blood loss, further randomized control study is needed which compare LADG and TLDG.

We compared WBC count and serum CRP level to evaluate the inflammatory response between the LADG and TLDG groups. Previous studies have reported lower serum CRP level in TLDG than that in LADG groups on postoperative day 7 [1112]. Although the level of inflammatory parameters on one certain day might reflect the inflammatory response, we thought that a decreasing tendency in these parameters would be more useful than the level itself. Thus, we analyzed the perioperative changes in WBC counts and serum CRP levels using a repeated-measures two-factor analysis, as shown in Table 3. WBC count and serum CRP level in the TLDG group were lower than those in the LADG group. The decreasing rate in WBC count in the TLDG group was also lower, whereas the decreasing rate in serum CRP level was not significantly different. Serum CRP usually reaches a peak later than that of WBCs [20]. Unfortunately, we only measured serum CRP until postoperative day 5, and different results could be expected if we measured it for a longer period. Early postoperative complication such as intra-abdominal abscess and wound infection can cause substantial changes in the inflammatory parameters. Therefore, we performed subgroup analysis for the patients without any complication, and there was similar result that WBC count (P = 0.012), serum CRP level (P = 0.034), and decrease rate of WBC count (P < 0.001) in the TLDG group were significantly lower than those in the LADG group.

Lower inflammatory response in TLDG might be caused by less tissue damage. Pulling the stomach through a minilaparotomy during LADG can cause adverse effects. In addition, excess traction at the laparotomy site is often needed during LADG because manipulation through the narrow field of minilaparotomy is sometimes very difficult. For example, LADG in obese patients is more difficult and can destroy more abdominal wall tissue. Many researchers have reported that TLDG could improve early surgical outcomes for obese patients [2122].

In this study, no difference in operative time was observed between the LADG and TLDG groups. However, operative time of TLDG has changed over time at our institute. As shown in Fig. 3, operative time decreased markedly after the eighth TLDG case. This change is thought to be due to shortening of anastomosis time and overcoming the learning curve for intracorporeal anastomosis. Advancements in other procedures (such as lymph node dissection without using the epigastric port for liver traction, use of a laparoscopic pouch for specimens, and taking the specimen out through the umbilical incision) also helped to reduce operative time. Although previous studies reported that 20-40 cases are needed to overcome the TLDG learning curve [2324], we experienced a shorter period to overcome.

There are several limitations in this study. First, the advancement of operative team and surgeon's learning curve might be factors that caused difference of outcome between LADG and TLDG groups. Although PSM was used for reducing selection bias, difference of operative period between two groups is major limitation of our study. However, all of operations in this study were performed by single surgeon who had experienced more than 80 cases of laparoscopic gastrectomy. Previous studies suggested that 50 laparoscopic gastrectomies are needed for overcome the learning curve [2526]. Therefore, the advancement of operative team and surgeon's learning curve is less likely to influence the results of our study. Second, we had studied the laboratory findings for acute inflammatory response as one of the short-term outcomes and there were statistically significant differences between the two groups. Because these parameters are not specific and can be influenced by many factors and various situations, its clinical implication is unclear. Finally, because this study is retrospective study focused on short-term operative outcome, we could not assess some important issues, like different reconstruction types, quality of life, and cost effectiveness between two groups. Also, the number of enrolled patients in the present study was too small to conduct subgroup analyses based on the reconstruction types.

In conclusion, this study suggests that the TLDG is one of the safe and feasible procedure for the treatment of early gastric cancer with the advantages in hospital stay, blood loss, and inflammatory response compared with LADG. However, well designed large scaled prospective randomized study is needed for proving the real benefits of TLDG.