All RCTs and non-RCTs comparing LGD2 with OGD2 for AGC were identified by searching the PubMed, EMBASE, and Cochrane Library databases for studies published between 1 January 2000 and 12 January 2013. Only articles published in English or Chinese were included in this study. The following medical subject headings and free-text terms were used: stomach neoplasms; stomach cancer; gastric carcinoma; gastric cancer; laparoscopy; laparoscopic; minimally invasive; laparotomy; conventional gastrectomy; OG; D2 lymph node dissection; extended; radical. Additional relevant articles were identified using references of relevant articles and previous meta-analyses. The PubMed database was used to search for additional studies and trials using authors’ names and the “related articles” function. The World Health Organization International Clinical Trials Registry Platform, Clinical Trials, Cochrane Central Register of Controlled Trials, and Chinese Clinical Trial Register were used to identify any ongoing RCTs.

Based on the preoperative clinical assessment or postoperative pathologic examination, AGC was defined as cancerous growth invading beyond the submucosal layer of the stomach. Locally AGC is the subgroup of AGC excluding stage IV. LG was defined as total LG or laparoscopy-assisted gastrectomy. In all included studies, D2 lymph node dissection was performed according to the JGCA lymph node classification[38]. The evaluated endpoints were classified as operative outcomes (operative time, intraoperative blood loss, and conversion rate), postoperative outcomes (postoperative analgesic consumption, time to first ambulation, time to first flatus, time to first oral intake, length of postoperative hospital stay, postoperative morbidity, incidence of reoperation, and postoperative mortality), and oncologic outcomes (number of lymph nodes harvested, tumor recurrence and metastasis, and disease-free and overall survival rates). The primary endpoints were postoperative morbidity and mortality as well as disease-free and overall survival rates. Other variables were considered as secondary endpoints.

The analyses included studies comparing LGD2 with OGD2 in patients with AGC. In cases when more than one publication reported on a single trial, only the most recent data were included, unless relevant outcomes were reported only in earlier publications. The following criteria were applied to exclude a study: < 40 cases; combined examination of AGC and EGC cases and/or D1-D3 lymphadenectomy, which prevented extraction of relevant or the authors’ provision of such data by email; malignant stromal tumors, benign disease, or emergency operations; use of hand-assisted LG, gasless laparoscopic surgery, or robotic surgery.

The meta-analyses were performed in accordance with the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-analysis statement[39]. Two reviewers (Zou ZH and Zhao LY) independently evaluated all retrieved studies to determine if they met the criteria, to assess study quality, and extract data. The study team resolved all of their disagreements through discussion to reach a consensus.

The Cochrane Collaboration Handbook 5.1.0 was used to independently determine the quality and risk of bias of RCTs[40]. Following domains were assessed: sequence generation; allocation concealment; completeness of outcome data; selective outcome reporting; baseline comparability of groups; dropout rates. The risk of bias in each domain was assessed and classified as low, high, or unclear. Blinding methods were not examined in this review because both LGD2 and OGD2 are invasive, and the patients were informed preoperatively about the planned procedures.

The methodological quality of non-RCTs was assessed using the modified Newcastle-Ottawa scale[41]. Patient selection, comparability of LGD2 and OGD2 groups, and assessment of measured outcomes were examined. In assessing comparability between groups, focus was on the variables that might affect primary endpoints such as, patient age and sex, pathologic tumor-node-metastasis stage, type of gastrectomy, resection margin, tumor size, histologic type, reconstruction, and adjuvant treatment.

Studies were scored using an ordinal star scale, with higher scores representing higher quality. A maximum of one star was awarded to a study for each numbered item within the selection and outcome assessment. A maximum of two stars was awarded for the comparability of the two groups. The quality of each study was graded as level 1 (0-5 stars) or level 2 (6-9 stars).

Review Manager (RevMan, version 5.0; The Cochrane Collaboration, Oxford, United Kingdom) and STATA (version 11.2; STATA Corporation, College Station, TX, United States) software were used for statistical analyses. Weighted mean differences (WMDs) with 95%CIs were calculated for continuous variables, including operative time, intraoperative blood loss, postoperative analgesic consumption, time to first ambulation, time to first flatus, time to first oral intake; length of postoperative hospital stay, and number of harvested lymph nodes. The odds ratios (ORs) with 95%CIs were calculated for dichotomous variables, including postoperative morbidity and mortality rates, incidence of reoperation, tumor recurrence, and metastasis. The hazard ratios (HRs) with 95%CIs extracted from Kaplan-Meier curves were used for disease-free and overall survival rates[42,43]. A random effects model was used to pool studies with significant heterogeneity, as determined by the χ² test (P≤ 0.10) and the inconsistency index (I²≥ 50%)[44,45].

An alternative statistical effect model was used to reanalyze the data for the sensitivity analysis (e.g., a random effects model instead of a fixed effects model or vice versa). The incidences of various postoperative morbidities and recurrence, and metastasis patterns were determined using subgroup analyses. The Begg’s test was used to assess the presence of publication bias. Publication bias was present when the continuity-corrected Pr >|z| value was ≤ 0.1[46].

Of the 493 publications identified in the initial literature search, 14 trials (1 RCT, 13 non-RCTs) were included in the analyses[19-32]. A total of 2596 participants (1328 in the LGD2 group and 1268 in the OGD2 group) were included in the study (Figure 1, Table 1).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Flow chart of the identification and inclusion of studies. Studies evaluating laparoscopic gastrectomy with D2 (LGD2) were identified to evaluate the procedure as an acceptable alternative to open gastrectomy with D2 (OGD2) for locally advanced gastric cancer (AGC); EGC: Early gastric cancer; RCT: Randomized controlled trial.

A methodological quality assessment revealed that the RCT had unclear random sequence generation, satisfactory allocation concealment, adequately addressed incomplete outcome data, and had no selective outcome reporting[23]. The quality of all 13 non-RCTs was level 2 (6-9 stars) on the Newcastle-Ottawa scale (Table 1).

Thirteen studies provided operative time data[19-25,27-32]. The LGD2 group’s weighted mean operative time was 57.06 min longer than in the OGD2 group (95%CI: 41.87-72.25; P < 0.01), with significant heterogeneity among studies (I² = 90%; P < 0.01) (Table 2, Figure 2A).

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Meta-analyses of procedure characteristics. A: Weighted mean operative time; B: Intraoperative blood loss. LGD2: Laparoscopic gastrectomy with D2 extended lymph node dissection; OGD2: Open gastrectomy with D2 extended lymph node dissection; RCT: Randomized controlled trial.

Blood loss data was found in 11 studies[19-23,25,27-30,32], revealing a significantly lower blood loss in the LGD2 compared to the OGD2 groups (WMD = -137.87 mL, 95%CI: -164.41--111.33; P < 0.01), with significant heterogeneity among studies (I² = 90%; P < 0.01) (Figure 2B).

Laparoscopic procedure conversion rates were documented in eight studies, ranging from 0.00 to 6.67%, with a weighted average of 1.68%[19,21-24,28,30,32]. Four articles reported the following reasons for converting to open procedures: hemorrhage (n = 2); overlarge tumor (n = 2); common bile duct injury (n = 1); obesity (n = 1); technical difficulty (n = 1); lack of pneumoperitoneum (n = 1); failure of the linear stapler (n = 1); dense adhesion after open sigmoidectomy (n = 1); relatively fixed tumor (n = 1); small incision metastasis (n = 1).

Analgesic administration was reported by only four articles included in this study[21,22,24,25]. Meta-analysis revealed a significantly lower frequency of analgesic administration in the LGD2 group than in the OGD2 group (WMD = -1.94, 95%CI: -2.50--1.38; P < 0.01), with significant heterogeneity among studies (I² = 77%; P < 0.01) (Table 2, Figure 3A).

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Meta-analyses of patient characteristics. A: Analgesic consumption; B: Time to first ambulation; C: Time to first flatus; D: Time to first oral intake. LGD2: Laparoscopic gastrectomy with D2 extended lymph node dissection; OGD2: Open gastrectomy with D2 extended lymph node dissection; RCT: Randomized controlled trial.

The time to first ambulation was reported in five papers[21,23,24,27,32]. This time was significantly shorter in the LGD2 group than in the OGD2 group (WMD = -1.03 d, 95%CI: -1.90--0.16; P < 0.05), with significant heterogeneity among studies (I² = 97%; P < 0.01) (Figure 3B).

The time to first flatus was reported in nine articles[19-24,27,30,31]. The time was significantly shorter in the LGD2 group than in the OGD2 group (WMD = -0.98 d, 95%CI: -1.30--0.66; P < 0.01), with significant heterogeneity among studies (I² = 89%; P < 0.01) (Figure 3C).

The time to first oral intake was reported in six papers[19,22-24,27,32]. Meta-analysis demonstrated this time was significantly shorter in the LGD2 group than in the OGD2 group (WMD = -0.85 d, 95%CI: -1.67--0.03; P < 0.05), with significant heterogeneity among studies (I² = 86%; P < 0.01) (Figure 3D).

The length of postoperative hospitalization was reported in 10 articles[19,20,22-25,27,28,30,32]. The LGD2 group had significantly shorter postoperative hospitalization than the OGD2 group (WMD = -3.08 d, 95%CI: -4.38--1.78; P < 0.01). There was a significant heterogeneity among studies (I² = 88%; P < 0.01) (Figure 4A).

Open in New Tab   Full Size Figure   Download Figure

Figure 4 Meta-analyses of postoperative events. A: Postoperative hospitalization; B: Morbidity; C: Reoperation; D: Mortality. LGD2: Laparoscopic gastrectomy with D2 extended lymph node dissection; OGD2: Open gastrectomy with D2 extended lymph node dissection; RCT: Randomized controlled trial.

The postoperative morbidity rates were reported in 13 studies[19-25,27-32]. Meta-analysis demonstrated a significantly lower overall postoperative morbidity after LGD2 than after OGD2 (OR = 0.78, 95%CI: 0.61-0.99; P < 0.05), with no significant heterogeneity among studies (I² = 14%) (Figure 4B).

The subgroup analyses showed significantly lower incidence rates of wound problems (wound infection and dehiscence) and pneumonia in the LGD2 group. No difference in the incidence rate of major surgical site complications, such as anastomotic stenosis, anastomotic leakage, duodenal stump leakage, pancreatic fistula or pancreatitis, and intra-abdominal bleeding, was found between the two groups (Table 2). Subgroup analyses demonstrated no significant differences between groups in major surgical site complications with regard to surgical extensions (distal gastrectomy/proximal gastrectomy/total gastrectomy). This includes anastomotic stenosis, anastomotic leakage, duodenal stump leakage, pancreatic fistula or pancreatitis, and intra-abdominal bleeding.

The reoperation incidence rate was reported in seven articles[19,20,22,24,30-32]. No significant difference in this parameter was found between the LGD2 and OGD2 groups (OR = 1.58, 95%CI: 0.58-4.31) with no significant heterogeneity among studies (I² = 0%) (Figure 4C).

The postoperative mortality rates were reported in 13 studies[19-25,27-32] with no significant difference in the rate between the LGD2 and OGD2 groups (OR = 0.69, 95%CI: 0.21-2.26) with no significant heterogeneity among studies (I² = 0%) (Figure 4D).

The number of lymph nodes harvested was reported in 13 studies[19-24,26-32]. Although meta-analysis showed no significant difference in this parameter between the two groups (WMD = -0.11, 95%CI: -2.72-2.50), there was significant heterogeneity among the studies (I² = 95%; P < 0.01) (Table 2, Figure 5A).

Open in New Tab   Full Size Figure   Download Figure

Figure 5 Meta-analyses of lymph node harvest and tumor recurrence. A: Lymph nodes harvested; B: Tumor recurrence and metastasis. LGD2: Laparoscopic gastrectomy with D2 extended lymph node dissection; OGD2: Open gastrectomy with D2 extended lymph node dissection; RCT: Randomized controlled trial.

Tumor recurrence and metastasis were recorded in eight studies[19-21,28-32]. The meta-analysis showed no statistical difference between the LGD2 and OGD2 groups (OR = 0.79, 95%CI: 0.60-1.04), as well as no significant heterogeneity among studies (I² = 20%) (Figure 5B). Subgroup analyses showed no significant difference in recurrence and metastasis patterns between the groups (Table 2).

Four trials involving 703 patients provided three-year disease-free survival rates[19,28,31,32]. Three trials involving 652 patients provided five-year disease-free survival rates[28,31,32]. The two groups showed no significant difference in three-year (HR = 1.02, 95%CI: 0.64-1.61) (Figure 6A) or five-year (HR = 1.02, 95%CI: 0.66-1.57) (Figure 6B) disease-free survival rates (Figure 6). There was no significant heterogeneity among studies (I² = 0% for both rates) (Table 2).

Open in New Tab   Full Size Figure   Download Figure

Figure 6 Meta-analyses of treatment outcomes between laparoscopic and open D2 gastrectomy. A: Three-year disease-free survival; B: Five-year disease-free survival; C: Three-year overall survival; D: Five-year overall survival. LGD2: Laparoscopic gastrectomy with D2 extended lymph node dissection; OGD2: Open gastrectomy with D2 extended lymph node dissection.

Eight trials involving 1363 patients provided three-year overall survival rates[19,23-25,27,28,31,32], and three trials involving 652 patients provided five-year overall survival rates[28,31,32]. The two groups showed no significant differences in three-year (HR = 0.87, 95%CI: 0.59-1.27) (Figure 6C) or five-year (HR = 0.79, 95%CI: 0.46-1.34) (Figure 6D) overall survival rates, accompanied with no significant heterogeneity among studies (I² = 0% for both). Among the studies included, only Shinohara et al[32] presented calculated disease-free and overall survival rates after LGD2 and OGD2 with regard to tumor stage, with no significant differences observed between the two groups.

Study results were reanalyzed using alternative (random or fixed effects) models showing no significant difference in pooled effects, except comparable incidences of overall morbidity in the two groups (OR = 0.78, 95%CI: 0.59-1.02). Furthermore, the studies showed no significant heterogeneity (I² = 14%). Endpoint analysis revealed no strong evidence of bias (Begg’s rank correlation test, continuity-corrected Pr >|z| > 0.1), except for tumor recurrence/metastasis (Pr >|z| = 0.035) and distant metastasis (Pr >|z| = 0.089).

Laparoscopic gastrectomy is gaining popularity worldwide as a minimally invasive alternative treatment to traditional open surgery in treating gastric cancer. The Japanese Gastric Cancer Association guidelines stipulate that D2 gastrectomy is required to cure advanced gastric cancer. However, the application of laparoscopic D2 gastrectomy (LGD2) remains questionable due to its technical difficulty and the lack of long-term results.

The authors performed a meta-analysis comparing LGD2 with open D2 gastrectomy (OGD2) in patients with advanced gastric cancer, evaluating endpoints of operative, postoperative, and oncological outcomes.

Compared with OGD2, LGD2 is a safer and more effective method, with lower overall morbidity, enhanced postoperative recovery, and comparable oncologic outcomes.

LGD2 is safe and effective, and offers some advantages over OGD2 in treatment of locally advanced gastric cancer. However, well-designed, prospective, multicenter, randomized controlled trials comparing LGD2 with OGD2 for treatment of advanced gastric cancer are warranted before recommending LGD2 for wider use in surgical practice.

The paper is a well-organized and structured meta-analysis of currently available data on the benefits of using LGD2 over OGD2 for the treatment of advanced gastric cancer. The authors concluded that LGD2 is safe and effective in treating locally advanced gastric cancer. This meta-analysis is well written and an important addition of knowledge to successful treatment of locally advanced gastric cancer.