Short‐term outcomes of laparoscopic and robotic distal gastrectomy for gastric cancer: Real‐world evidence from a large‐scale inpatient database in Japan

Robotic distal gastrectomy (RDG) has been widely performed throughout Japan since it became insured in 2018. This study aimed to evaluate the short‐term outcomes of RDG and laparoscopic distal gastrectomy (LDG) for gastric cancer using real‐world data.

Gastric cancer is one of the most common malignant tumors in far-eastern countries.Moreover, it is the fifth most common malignancy and third leading cause of cancer-related death worldwide. 1[7] Robotic surgical systems have recently been developed to overcome these limitations.][10][11] Some studies have indicated that robotic distal gastrectomy (RDG) was associated with better short-term outcomes in terms of postoperative complications 10,[12][13][14] and length of hospital stay (LOS) 15,16 compared with LDG.However, these data were based on a single-center study or clinical trials involving experienced surgeons and institutions and may not be applicable to the general medical community.As an illustrative example, there was a difference in surgical outcomes between patients who underwent open and laparoscopic gastrectomy at centers of excellence and those in real clinical settings. 17g data from the National Clinical Database in Japan between October 2018 and December 2019 showed comparable outcomes between RDG and LDG. 18In that study, the time elapsed between April 2018, when robotic surgery became covered by insurance, and the data collection period was short.
In addition, it is possible that few new facilities introduced RDG during this period owing to multiple restrictions on initiating RDG at inexperienced facilities. 19,20Based on these factors, the results of the previous study 18 may differ from current real-world outcomes, in which RDG has been widely performed throughout Japan.
In this study, we compared the short-term surgical outcomes between patients treated with RDG and LDG using a large-scale database in Japan to reveal the real-world impact of RDG for gastric cancer.

| Data source
There are 47 prefectures in Japan.Furthermore, there is at least one public or private university hospital in each prefecture that provides advanced medical care in their respective areas.All 42 national university hospitals, which are teaching hospitals, were selected as representatives of the whole of Japan (Figure 1).This   comorbidities was converted into a score that was summed up for each patient. 22Postoperative complications were identified during hospitalization.The study was approved by the Institutional Ethical Committee of Shinshu University (approval number: 5740).The requirement for informed consent was waived due to the anonymous nature of the data.

| Statistical analysis
Propensity-score matching of variables in Table 1 was not performed in the present study due to low c-statistic for goodness-of-fit test (0.500 [95% confidence interval (CI), 0.474-0.526]).Categorical variables were compared using the χ 2 test.Continuous variables are presented as medians with interquartile ranges (IQRs) and were compared using the Mann-Whitney U test.Overall postoperative complications and in-hospital mortality were assessed using logistic regression analysis.The results of the logistic regression analysis are presented as odds ratios (ORs) with 95% CIs.Time to diet resumption and postoperative LOS were assessed using Cox regression analysis.The estimates of the Cox regression analysis are presented as hazard ratios (HRs) with 95% CIs.All reported p values were two-tailed.Additionally, a p value of <0.05 was considered statistically significant.Stata/MP 15.0 (Stata Corp.) was used for the statistical analysis.

| RESULTS
A total of 46,261 patients with gastric cancer were admitted to all 42 hospitals between April 2018 and October 2022 (Figure 2).Among them, 4161 patients who underwent LDG (n = 3173) or RDG (n = 988) were selected.
Table 1 presents the patient demographics in the two groups.
Age, BMI, Charlson comorbidity index, and clinical cancer stage differed between the two groups.More patients in the RDG group were diagnosed with clinical stage ⅠB-Ⅲ gastric cancer than those in the LDG group (43.3% vs. 35.9%,p < 0.001).
There was no difference in in-hospital mortality, in-hospital mortality within 30 days of surgery, or postoperative complication rates between the two groups (Table 2).The logistic regression analysis showed that older age, 70-79 years (OR p < 0.001) compared with ≤2 were associated with increased incidence of overall postoperative complications (Table 3).By contrast, female sex was associated with decreased incidence of overall postoperative complications (OR 0.56 [IQR 0.42-0.75],p < 0.001).RDG did not increase the risk of complications compared with LDG (Table 4).Using Cox regression analysis, RDG was associated with shorter time to diet resumption after surgery (HR 1.18, 95% CI [1.01-1.27],p < 0.001) (Supporting Information: Table 1) and LOS (HR 1.14, 95% CI [1.06-1.23],p < 0.001) compared with LDG (Supporting Information: Table 2).

| DISCUSSION
The present study revealed that RDG was associated with low inhospital mortality and postoperative complication rates comparable with those of LDG.Subsequent results showed that RDG increased the duration of anesthesia and was associated with comparable time to diet resumption and shorter postoperative LOS.Mortalities in LDG and RDG were 0.1-0.5% and 0-0.6%, respectively.Moreover, morbidities were 3.9%-18.4%and 1.0%-18.4%,respectively. 12,18,23,24The low mortality rates in the present study are equivalent to those in previous studies, suggesting that high-quality LDG and RDG were performed in the real world throughout Japan over the 4 years since RDG became covered by insurance.
Several randomized control trials (RCTs) have revealed no significant differences in morbidity between LDG and RDG. 10,15,16By contrast, one RCT 13 and some nonrandomized studies 9,14 reported a lower incidence of postoperative complications in RDG compared with LDG.One of the advantages of RDG is that it is a robotic approach that uses an articulating arm.
6][27] and 0%-2.5%, 12,13,25,27spectively.0][31] An articulating arm can be a solution for pancreatic damage that occurs during lymph node dissection of the suprapancreatic region. 25Based on the results of the present and recent other studies, determining whether LDG or RDG should be performed in a particular patient is difficult.However, evaluating the efficacy of robotic surgery is important.Particularly, assessing whether robotic surgery can decrease the difficulty in technically demanding procedures such as radical resection after neoadjuvant chemotherapy, other types of gastric surgery namely total and proximal gastrectomy, and transhiatal approach for esophagogastric junction cancer is important.
Flowchart of the study group selection.
RDG has been reported to be associated with a longer procedure duration 10,13,15,18 with shorter postoperative LOS than did LDG, 12,15,16,18 which are in line with the results of the present study.This indicates that a longer procedure duration in RDG does not affect its invasiveness.By contrast, Omori et al. 24 reported a shorter operative time in RDG than in LDG (208 vs. 231 min, p < 0.005).Because the learning curve of RDG is 10-25 cases, [32][33][34] once surgeons become experienced in these cases, the duration can be reduced.
In the present study, propensity-score matching was not performed due to low c-statistic for goodness-of-fit test (0.5).This low value indicated that many pairs of LDG and RDG cases showing a close propensity score could be formed.
Therefore, a multivariate analysis was performed without propensity-score matching because it would reduce the number of cases.
However, variables in which statistically significant differences existed between the two groups, such as age, BMI, Barthel Index, Charlson comorbidity index, and clinical cancer stage (Table 1) might be potential confounders.Particularly, these confounders were considered when comparing clinical outcomes, namely time to diet resumption and postoperative LOS, between both groups.Adjusting for possible confounding bias, Cox regression analyses were performed, from which the estimates are presented in Supporting Information: Tables 1 and 2. As a result, RDG was independently p < 0.001) (Table 5).This paradox was due to the different distributions indicated by the differences in the IQRs (2-3 and 3-4 days) and the method used to compare continuous variables: the Postoperative complications, in-hospital mortality, and 30-day unplanned readmission after hospital discharge in laparoscopic and robotic distal gastrectomy.However, time to diet resumption was considered similar between the two groups, because the median value was the same.
A key strength of the present study is that the data were collected from 42 national university hospitals dispersed throughout Japan.This means that the data were representative samples from all over Japan.In addition, since the data were collected in 2018 and later, a recent and short period of time, this represents the current quality of surgery.
The present study has several limitations owing to the nature of the database.First, the definition of complications was not standardized.Additionally, information on severity grades, specifically the Clavien-Dindo (C-D) classification, was lacking.This is because the DPC data is an administrative claims database, such as the Nationwide Inpatient Sample in the United States. 35 This large-scale study revealed that RDG was performed safely and provided shorter postoperative LOS nationwide, including in regional cities, since it became covered by insurance in Japan.
study used data from the Japanese Diagnosis Procedure Combination (DPC) database of these hospitals.Data of patients with stage Ⅰ, Ⅱ, or Ⅲ gastric cancer who underwent LDG or RDG between April 2018 and October 2022 were collected.The database includes administrative claims information and discharge; diagnosis; patient demographics; comorbidities upon admission, complications after admission using the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) codes and procedures; surgeries performed (coded with original Japanese codes); duration of anesthesia; and use of blood transfusions.

2. 2 |
Patient selection and data Patients aged >19 years who were admitted to hospitals with a precipitating diagnosis of gastric cancer (ICD-10 code C16) from April 1, 2018, to October 31, 2022, in 42 national university hospitals and underwent LDG or RDG were included.A 3-month look back window before each patient's selection into the study cohort was used for assessing the presence of chemotherapy before surgery for gastric cancer.Body mass index (BMI) was classified into four categories (<18.5, 18.5-25, 25-35, and >35 kg/m 2 ).The Charlson comorbidity index was calculated based on Quan's protocol, and each ICD-10 code for 12
associated with decreased time to diet resumption and postoperative LOS.Time to diet resumption in the RDG group was significantly shorter than that in the LDG group, despite having the same median number in both groups (3 days [IQR, 2-3] vs. 3 days [IQR, 3-4],

Table 5
Characteristics of patients who received laparoscopic or robotic distal gastrectomy.
lists the duration of anesthesia, postoperative LOS, time to diet resumption, and total costs in the two groups.RDG T A B L E 1 Abbreviation: BMI, body mass index.
Logistic regression analysis of postoperative complications.Logistic regression analysis of each specific complication of robotic gastrectomy with reference to laparoscopic gastrectomy.
outcomes could not be assessed.Finally, information on resected specimens, such as the number of lymph nodes dissected, and margin status is lacking.Therefore, comparing the quality of surgery between RDG and LDG from that perspective is impossible.T A B L E 3 T A B L E 4Abbreviations: CI, confidence interval; Inf, infinity; NA, not available.