Comparison of long-term quality of life between Billroth-I and Roux-en-Y anastomosis after distal gastrectomy for gastric cancer: a randomized controlled trial

Abstract

Background:

The results of studies comparing Billroth-I (B-I) with Roux-en-Y (R-Y) reconstruction on the quality of life (QoL) are still inconsistent. The aim of this trial was to compare the long-term QoL of B-I with R-Y anastomosis after curative distal gastrectomy for gastric cancer.

Methods:

A total of 140 patients undergoing curative distal gastrectomy with D2 lymphadenectomy in West China Hospital, Sichuan University from May 2011 to May 2014 were randomly assigned to the B-I group (N = 70) and R-Y group (N = 70). The follow-up time points were 1, 3, 6, 9, 12, 24, 36, 48, and 60 months after the operation. The final follow-up time was May 2019. The clinicopathological features, operative safety, postoperative recovery, long-term survival as well as QoL were compared, among which QoL score was the primary outcome. An intention-to-treat analysis was applied.

Results:

The baseline characteristics were comparable between the two groups. There were no statistically significant differences in terms of postoperative morbidity and mortality rates, and postoperative recovery between the two groups. Less estimated blood loss and shorter surgical duration were found in the B-I group. There were no statistically significant differences in 5-year overall survival (79% [55/70] of the B-I group vs. 80% [56/70] of the R-Y group, P = 0.966) and recurrence-free survival rates (79% [55/70] of the B-I group vs. 78% [55/70] of the R-Y group, P = 0.979) between the two groups. The scores of the global health status of the R-Y group were higher than those of the B-I group with statistically significant differences (postoperative 1 year: 85.4 ± 13.1 vs. 88.8 ± 16.1, P = 0.033; postoperative 3 year: 87.3 ± 15.2 vs. 92.8 ± 11.3, P = 0.028; postoperative 5 year: 90.9 ± 13.7 vs. 96.4 ± 5.6, P = 0.010), and the reflux (postoperative 3 year: 8.8 ± 12.9 vs. 2.8 ± 5.3, P = 0.001; postoperative 5 year: 5.1 ± 9.8 vs. 1.8 ± 4.7, P = 0.033) and epigastric pain (postoperative 1 year: 11.8 ± 12.7 vs. 6.1 ± 8.8, P = 0.008; postoperative 3 year: 9.4 ± 10.6 vs. 4.6 ± 7.9, P = 0.006; postoperative 5 year: 6.0 ± 8.9 vs. 2.7 ± 4.6, P = 0.022) were milder in the R-Y group than those of the B-I group at the postoperative 1, 3, and 5-year time points.

Conclusions:

Compared with B-I group, R-Y reconstruction was associated with better long-term QoL by reducing reflux and epigastric pain, without changing survival outcomes.

Trial Registration:

ChiCTR.org.cn, ChiCTR-TRC-10001434.

Introduction

Gastric cancer is one of the most common cancers in the world and the leading cause of cancer-related deaths worldwide, especially in East Asia.^[1] Surgery is still the most important component of treatments for gastric cancer. However, patients undergoing gastrectomy may experience a variety of symptoms, including abnormal diet, weight loss, and appetite loss, which can lead to a significant deterioration of the patient's quality of life (QoL).^[2,3] In recent years, the survival of patients with gastric cancer has been significantly improved due to increasing early detection and treatment, the performance of standard D2 gastrectomy, and progresses of comprehensive therapies.^[4–6] Therefore, the QoL and type of digestive tract reconstruction which could greatly influence QoL have been paid more attentions to than before.

Billroth-I (B-I) and Roux-en-Y (R-Y) reconstructions are traditional reconstruction approaches after distal gastrectomy. B-I reconstruction has the physiological advantage that food can pass directly into the duodenum through the remnant stomach. However, many patients may experience obvious symptoms, including upper abdominal pain and indigestion; while duodenogastric reflux has been considered as the main cause of clinical symptoms after B-I reconstruction.^[7] In comparison, the R-Y reconstruction has the merits of prevention from bile reflux.^[8] R-Y reconstruction has the disadvantages of technically complexity, stasis syndrome, and difficulty of postoperative endoscopy examination yet.^[9,10]

Several prospective randomized controlled clinical trials (RCTs) have compared morbidity, mortality, and QoL between B-I and R-Y anastomosis after distal gastrectomy.^{[7–9,11–13]} Their findings, however, are still inconsistent. Furthermore, most of these RCTs assessed the short-term QoL of patients with relatively early cancer within 1 year after the surgery. Long-term dynamically assessments of the postoperative QoL of patients with relatively advanced stage at multiple time points were rare. Furthermore, the impact of different digestive tract constructions after distal gastrectomy on the long-term survival also needs to be addressed. To resolve these issues, we have performed an RCT to compare B-I with R-Y reconstruction after curative distal gastrectomy for gastric cancer in China, and found that the stronger anti-reflux capability of R-Y anastomosis contributed to a higher QoL by reducing the reflux-related gastritis and pain symptoms, and promoted a better global health 1 year after the surgery.^[9] Here, we present the results of long-term QoL at 36 and 60 months postoperatively.

Methods

Patients and ethical approval

The inclusion criteria have been reported and published previously,^[9] and listed briefly as follows: patients aged 18 to 75 years and surgically tolerant; the World Health Organization performance status score was <2; gastric adenocarcinoma was diagnosed by preoperative gastroscopy or biopsy; the tumor was located in the lower third of the stomach, and distal gastrectomy including open and laparoscopic-assisted gastrectomy was planned to be performed; the preoperative staging of resectable tumors was less than T4aN2M0 according to the 3^rd English version of Japanese classification of Gastric Cancer.^[14] Exclusion criteria included: patients with a history of laparotomy (except for appendectomy and laparoscopic cholecystectomy); patients who required total gastrectomy or combined organ resection (except for cholecystectomy) for treatment; patients were diagnosed with other gastric malignancies (such as gastrointestinal stromal tumor), or any previous malignancies or synchronous malignancies; emergency cases due to tumor perforation or bleeding; patients received neoadjuvant chemotherapy or perioperative radiotherapy. This study was registered on the Chinese Clinical Trial Register (ChiCTR), No. ChiCTR-TRC-10001434, and approved by the West China Hospital of Sichuan University, Biomedical Research Ethics Committee (No. 2010[54]), and the study was done in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients.

Study design

This study was designed as a single-center randomized controlled trial. This study used a simple random sampling method. The random allocation sequence was generated from the random number table, then it was hidden and sealed in opaque and sequential numbered envelopes until the patients’ eligibilities were confirmed after the initial intraoperative exploration. Subsequently, the included patients were randomly divided into the B-I group and R-Y group at a ratio of 1:1 during the operation. In this study, the patients, surgeons, and staff who collected data and analyzed the results were not blinded. The research protocol has not been revised during the research period.

Sample size

The sample size was calculated by using a two-sided alpha error of 0.05 under the normal distribution with a statistical power of 0.9 (β = 0.1). And the effect size was determined based on the reported data which showed that the heartburn rates of the B-I group and R-Y group were 37% and 8%, respectively, 1 year after the operation.^[15] Consequently, the planned sample size was calculated as 56 in each arm. Finally, 70 patients were enrolled in each arm, considering a dropout rate of 15% to 20%.

Surgical technique and postoperative care

The surgical techniques had been described in our previous article.^[9] Briefly, all the patients underwent distal gastrectomy with D2 lymphadenectomy in accordance with the Japanese gastric cancer treatment guidelines (version 3.0).^[16] All anastomoses were completed by using mechanical staplers and reinforced with intermittent full-thickness sutures. In the B-I group, a 25-mm circular stapler was used to perform an end-to-side gastroduodenal anastomosis between the duodenal stump and the posterior wall of the remnant stomach. And the stump of the remnant stomach was closed with a linear stapler. In the R-Y group, the duodenum was transected and closed with a linear stapler at 3 cm distal to the pylorus, with reinforcement by interrupted full-thickness sutures and interrupted seromuscular sutures. During reconstruction, the proximal jejunum was identified and transected 15 to 20 cm from the Treitz ligament firstly. Then, a 25-mm circular stapler was used to perform a side-to-side gastrojejunostomy between the posterior wall of the remnant stomach and the antimesenteric border of the distal jejunum. Next, a side-to-side jejunojejunostomy was established between the antimesenteric borders of the proximal jejunum and the distal jejunum 45 cm below the gastrojejunostomy. Linear staplers were used to close the stumps of the remnant stomach and two jejunums. The mesenteric defect was sutured interruptedly. The research team evaluated surgical compliance and surgical quality by reviewing intraoperative photos. After the operation, a standardized postoperative care plan was adopted and has been described previously.^[9]

Follow-up

The follow-up time points were 1, 3, 6, 9, 12, 24, 36, 48, and 60 months after the operation. The final follow-up time was May 2019. Information on patients’ survival, recurrence or metastasis status, adjuvant therapy, and questionnaires were collected. Meanwhile, pieces of advice were given to the patients regarding the treatment and lifestyle during the follow-up. The follow-up methods included hospital out-patients follow-up, telephone follow-up, online follow-up, letter follow-up, etc. For educated patients, when they came to the out-patient clinic, we asked them to fill out the questionnaires themselves. For mail follow-up, we enclosed the questionnaires in an envelope and asked them to fill and sent it back. During the online or telephone follow-up, we asked questions according to the contents of the questionnaires and filled them in. For uneducated patients, we asked the patients according to the questions of the questionnaires, and then the inquirer filled in the questionnaires.

Outcome measurements

The primary outcome was QoL. In this study, the data of QoL were collected by the validated Chinese version of the European Organization for Research and Treatment of Cancer (EORTC) core questionnaire (EORTC QLQ-C30, version 3.0), and stomach module questionnaire (EORTC QLQ-STO22).^[17–21] The EORTC QLQ-C30 questionnaire contains 30 items, which could be incorporated into five functional scales (physical, role, cognitive, emotional, and social), three symptom scales (pain, fatigue, nausea, and vomiting), six single items (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties), and one global health status scale. Only the item values of the global health scale range from 1 to 7; while other items have been divided into four response categories, namely “Not at all,” “A little,” “Quite a bit,” and “Very much,” corresponding to scores from 1 to 4.^[19] The EORTC QLQ-STO22 questionnaire includes 22 items, which could be divided into five symptom scales (dysphagia, epigastric pain, eating restrictions, reflux, symptoms, and anxiety) and four single items (having a dry mouth, taste, body image, and hair loss).^[18] A high score represents a high/healthy level of functioning or QoL for a functional scale or the global health status, or a high level of symptomatology/problems for symptom scales/items.^[19] All QLQ-C30 and QLQ-STO22 responses were linearly transformed to scores from 0 to 100 according to EORTC scoring manuals.^[19] The preoperative baseline of QoL was obtained 3 days before the operation. Postoperative QoL, recurrence, and survival status were collected every 3 months after the operation by telephone calls, letters, or outpatient visits. No patient was lost to follow-up. The secondary outcomes were operative safety, postoperative recovery, and severity of postoperative gastritis, which had been analyzed in our 1-year interim analysis.^[9]

Data collection

The data were collected by the research assistant using the Case Report Form firstly, and then the data in the Case Report Form was input into a predesigned EXCEL (Microsoft, Redmond, USA) electronic database for analysis. At the same time, each patient's medical record was scanned into a PDF version and stored for data traceability. Two staffs were appointed as the Clinical Research Associate to check the data regularly according to the original medical record and trial protocol.

Establishment of database

We used EXCEL to construct the electronic database. The table head was designed according to the Case Report Form. For digital data, we input it directly. For other types of data, we categorized the table head of items in advance, to transform the text data into digital data. In this way, it is convenient for data input, and avoided the need for re-conversion during data analysis. As for the quality control of data, we had a special scientific research assistant to check the input data according to the original data of patients to ensure that the input data was true and correct.

Statistical analyses

An intention-to-treat analysis was applied when analyzing the data. SPSS 25.0 software (SPSS, Chicago, IL, USA) was used for statistical analysis. Quantitative data were expressed as mean and standard deviation, and tested by a one-way analysis of variance (ANOVA) test if normal distributed. Otherwise, median and interquartile range would be used to describe the data. For categorical data, the chi-squared or Fisher's exact test was used to compare frequencies. Non-parametric tests were performed for non-normal distribution data. Survival analysis and curves were presented by the Kaplan–Meier analysis and compared by the log-rank test. A P value of ≤0.05 (two-sided) was considered statistically significant.

Results

Characteristics of patients

From May 2011 to May 2014, 140 patients with gastric cancer were included and allocated randomly into the B-I group (70 cases) and R-Y group (70 cases). Six patients in the R-Y group had to be changed to B-I reconstructions because of their willingness or economical consideration of patients’ families. In the B-I group, six patients converted to R-Y anastomosis due to the presence of tension between the remnant stomach and duodenum. The baseline characteristics of patients were comparable between the two groups.^[9]Figure 1 shows the trial profile.

Figure 1.

Flow diagram of patients in Billroth-I group and Roux-en-Y group to show the analyse process.

Operative variables and postoperative recovery

Operative details, including morbidity, mortality, and operative recovery parameters, have been reported previously.^[9] There were no significant differences regarding mortality and morbidity between the groups. Operative time was significantly shorter for B-I anastomosis (239.4 ± 40.8 min) than R-Y anastomosis (271.2 ± 39.2 min) (P < 0.001). The estimated blood loss was also significantly reduced in the B-I group (84.1 ± 32.0 mL vs. 104.2 ± 35.3 mL, P < 0.001). The early postoperative outcomes were comparable in terms of postoperative hospital stay, time to first gas-passing, time to first oral intake, time to nasogastric decompression removal, and time to drain removal between the two groups.^[9]

Long-term survival

Seventeen patients died within 3 years after the surgery, including ten in the B-I group and seven in the R-Y group. Twenty-nine patients, including 15 in the B-I group and 14 in the R-Y group died within 5 years after the surgery. Three years after surgery, there were 13 cases of recurrence in the B-I group and nine cases in the R-Y group. Totally 15 and 15 cases recurred within 5 years after the surgery in the B-I and R-Y groups, respectively. The overall 5-year survival rate of the B-I group was 79% (55/70), and 80% (56/70) of the R-Y group. There was no statistically significant difference in the overall survival rate between the two groups (χ² = 0.002, P = 0.966) [Figure 2A]. There was no statistically significant difference in the 5-year recurrence-free survival rate (79% [55/70] vs. 78% [55/70]) between the two groups (χ² = 0.001, P = 0.979) [Figure 2B].

Figure 2.

Survival analyses for patients between the B-I group and R-Y group. A: Overall survival rate; B: Recurrence-free survival rate. B-I: Billroth-I anastomosis; R-Y: and Roux-en-Y anastomosis.

Assessment of QoL

The preoperative QoL between the two groups were almost comparable, except for the insomnia item. There were no statistical differences between the two groups at 3 and 6 months after the surgery on the EORTC QLQ-C30 items. Poor insomnia was identified in the R-Y group with significant differences at postoperative 9 months, but turned to insignificant from postoperative 12 to 60 months. At the postoperative 1-, 3-, and 5-year time points, the scores of the global health status of the R-Y group were higher than those of the B-I group with statistical differences. A lower score of pain was observed in the R-Y group at the postoperative 1-year time point. However, no significant differences in terms of pain have been detected at postoperative 3 and 5 years. Except for the score of global health status, the other scales of QLQ-C30 items were not significantly different between the two groups at postoperative 5 years [Table 1 and Figure 3]. During the 5 years, the scores of physical functioning, role functioning, emotional functioning, social functioning, and global health status increased; while the scores of pain, fatigue, appetite loss, nausea, and vomiting showed decreased trends over time. Multivariate ANOVA did not show the interaction between time trends and grouping, which meant the variation of each scale over time trend was independent of grouping.

Table 1.

The preoperative and postoperative QoL of gastric cancer patients undergoing distal gastrectomy between B-I and R-Y group according to the EORTC QLQ-C30 items.

	Preoperative			Postoperative 3 months			Postoperative 6 months			Postoperative 9 months			Postoperative 12 months			Postoperative 36 months			Postoperative 60 months
Items	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	Pvalue
Physical functioning	93.1 ± 7.9	93.4 ± 8.3	0.710	92.4 ± 11.8	91.3 ± 10.8	0.430	95.2 ± 8.8	96.6 ± 6.9	0.356	93.9 ± 9.5	96.3 ± 8.8	0.073	96.7 ± 7.3	97.5 ± 7.5	0.339	98.0 ± 4.4	99.2 ± 2.2	0.058	98.9 ± 2.8	99.3 ± 2.0	0.381
Role functioning	93.4 ± 12.9	91.4 ± 16.7	0.966	90.4 ± 15.9	90.4 ± 15.3	0.916	94.4 ± 13.1	94.9 ± 12.6	0.929	92.4 ± 13.3	94.4 ± 11.4	0.459	96.3 ± 10.3	99.0 ± 4.0	0.124	97.4 ± 7.6	96.9 ± 10.4	0.802	98.5 ± 5.9	99.3 ± 3.3	0.367
Dyspnea	3.4 ± 10.2	4.4 ± 11.4	0.596	6.4 ± 16.5	2.9 ± 9.5	0.259	4.9 ± 15.5	2.5 ± 10.5	0.336	6.4 ± 14.4	3.9 ± 13.5	0.149	3.4 ± 11.7	0.5 ± 4.0	0.053	2.3 ± 8.6	0.6 ± 4.3	0.155	1.2 ± 6.4	0.7 ± 4.8	0.620
Pain	18.6 ± 20.5	21.8 ± 17.1	0.143	19.4 ± 25.0	13.2 ± 15.9	0.382	12.7 ± 18.5	7.6 ± 11.7	0.235	12.7 ± 19.1	6.6 ± 11.9	0.078	10.0 ± 13.9	4.2 ± 9.7	0.004	4.1 ± 9.1	2.5 ± 6.0	0262	3.4 ± 7.5	2.0 ± 5.5	0.304
Fatigue	17.0 ± 17.6	16.0 ± 16.6	0.814	13.2 ± 14.1	11.3 ± 13.4	0.401	10.5 ± 14.5	10.0 ± 14.1	0.757	11.3 ± 16.0	8.8 ± 13.5	0.340	9.3 ± 14.9	7.0 ± 13.3	0.319	4.1 ± 8.0	3.7 ± 8.6	0.801	2.1 ± 6.1	1.8 ± 5.2	0.830
Insomnia	19.6 ± 23.9	8.8 ± 15.9	0.004	8.8 ± 16.9	5.9 ± 14.0	0.277	9.8 ± 19.2	6.4 ± 16.5	0.253	12.3 ± 6.4	19.9 ± 16.5	0.031	8.3 ± 16.7	3.9 ± 12.3	0.064	4.1 ± 12.7	2.8 ± 9.3	0.522	2.5 ± 8.8	1.4 ± 6.7	0.477
Appetite loss	11.3 ± 22.0	12.3 ± 19.0	0.487	6.9 ± 13.6	9.3 ± 17.1	0.488	8.3 ± 16.7	4.9 ± 13.2	0.176	5.9 ± 14.0	4.4 ± 14.0	0.341	4.9 ± 13.2	2.5 ± 10.5	0.153	2.9 ± 9.5	2.8 ± 11.1	0.939	1.9 ± 7.7	0.7 ± 4.8	0.361
Nausea and vomiting	9.6 ± 19.8	10.3 ± 15.0	0.395	6.4 ± 12.2	6.1 ± 11.1	0.839	2.0 ± 6.1	3.7 ± 8.6	0.207	5.4 ± 15.1	4.4 ± 10.6	0.852	2.7 ± 8.9	2.0 ± 6.8	0.752	1.5 ± 5.7	1.7 ± 5.0	0.837	1.5 ± 5.9	0.3 ± 2.4	0.183
Constipation	7.8 ± 14.2	10.8 ± 18.6	0.470	6.9 ± 15.8	3.4 ± 11.7	0.131	3.4 ± 11.7	2.5 ± 8.8	0.737	4.4 ± 15.1	3.4 ± 10.2	0.841	2.9 ± 11.1	1.5 ± 6.9	0.458	1.8 ± 7.5	1.7 ± 7.3	0.949	1.9 ± 7.7	0.7 ± 4.8	0.361
Diarrhea	8.3 ± 15.6	11.8 ± 19.8	0.322	7.8 ± 17.4	11.8 ± 19.8	0.173	10.3 ± 21.7	9.8 ± 20.8	0.966	6.9 ± 14.7	6.4 ± 14.4	0.828	6.9 ± 14.7	7.8 ± 19.2	0.897	4.1 ± 12.7	2.2 ± 8.4	0.347	1.9 ± 7.7	0.7 ± 4.8	0.361
Cognitive functioning	89.2 ± 17.5	90.9 ± 13.0	0.795	96.8 ± 8.8	97.3 ± 7.9	0.932	96.1 ± 10.0	98.0 ± 7.4	0.121	95.3 ± 12.2	96.8 ± 10.1	0.448	97.1 ± 10.0	98.3 ± 7.1	0.379	99.1 ± 3.8	99.4 ± 3.0	0.610	99.7 ± 2.3	100.0 ± 0	0.343
Emotional functioning	83.9 ± 20.0	83.1 ± 15.4	0.319	96.2 ± 8.7	95.3 ± 10.2	0.773	96.2 ± 8.7	97.1 ± 7.6	0.613	94.7 ± 11.6	96.1 ± 9.7	0.502	95.2 ± 11.2	96.4 ± 8.6	0.857	97.4 ± 7.2	98.5 ± 4.2	0.313	98.0 ± 5.6	99.3 ± 2.9	0.139
Social functioning	86.3 ± 20.3	85.3 ± 19.9	0.799	95.3 ± 10.3	95.1 ± 12.6	0.621	96.3 ± 10.7	97.1 ± 10.8	0.572	96.8 ± 9.7	95.6 ± 14.0	0.767	98.0 ± 6.8	97.3 ± 12.4	0.557	99.1 ± 4.9	98.3 ± 5.0	0.392	99.4 ± 4.5	99.3 ± 3.3	0.936
Financial difficulties	22.1 ± 31.9	21.6 ± 28.1	0.822	5.9 ± 15.2	6.9 ± 16.8	0.673	4.9 ± 11.9	4.9 ± 15.5	0.641	5.4 ± 12.4	5.4 ± 16.9	0.510	8.8 ± 20.5	4.9 ± 16.6	0.149	4.1 ± 11.0	1.7 ± 7.3	0.162	1.2 ± 6.4	1.4 ± 6.7	0.922
Global Health status	66.3 ± 27.3	61.2 ± 23.8	0.119	73.4 ± 21.5	72.5 ± 17.9	0.531	80.0 ± 20.3	80.0 ± 17.8	0.760	80.6 ± 18.3	84.2 ± 18.3	0.174	85.4 ± 13.1	88.8 ± 16.1	0.033	87.3 ± 15.2	92.8 ± 11.3	0.028	90.9 ± 13.7	96.4 ± 5.6	0.010

B-I group: Billroth-I group; EORTC: European Organization for Research and Treatment of Cancer; QoL: Quality of life; R-Y group: Roux-en-Y group.

Figure 3.

Comparison of QoL scores from preoperative baseline to postoperative 5 year between the two groups. (A) Reflux symptom item in QLQ-STO22 questionnaire; (B) Global Health Status in QLQ-C30 questionnaire; (C) Epigastric Pain item in QLQ-STO22 questionnaire.

As for QLQ-STO22 items, the incidence of reflux symptoms after R-Y anastomosis was almost lower at each time point. And, there were significant differences on epigastric pain from postoperative 12 to 60 months. Three years after the operation, the items of having a dry mouth, anxiety, and hair loss in the R-Y group were significantly better than the B-I group. However, these items were fluctuant and turned to insignificant between the two groups at postoperative 5 years. There were no significant differences in terms of dysphagia, eating restrictions, taste, and body image at each time point [Table 2 and Figure 3]. Epigastric pain, reflux, and eating restrictions were improved from 3 months to 5 years after the operation. However, these trends have not been observed on other scales. Multivariate ANOVA also showed that there was no interaction between time trend and grouping.

Table 2.

The preoperative and postoperative QoL of gastric cancer patients undergoing distal gastrectomy between the B-I and R-Y group according to the EORTC QLQ-STO22 items.

	Preoperative			Postoperative 3 months			Postoperative 6 months			Postoperative 9 months			Postoperative 12 months			Postoperative 36 months			Postoperative 60 months
Items	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	P value	B-I group	R-Y group	P value
Dysphagia	4.7 ± 9.5	5.6 ± 10.2	0.736	6.4 ± 12.6	7.0 ± 11.2	0.529	4.2 ± 11.4	3.4 ± 7.5	0.993	6.4 ± 11.6	4.4 ± 9.6	0.481	3.1 ± 8.3	3.4 ± 9.6	0.861	2.9 ± 6.8	1.5 ± 5.6	0.212	1.4 ± 3.8	1.8 ± 5.2	0.677
Epigastric pain	21.4 ± 15.3	25.0 ± 14.0	0.160	20.8 ± 19.5	17.8 ± 16.5	0.348	16.4 ± 15.8	12.1 ± 12.9	0.105	15.4 ± 17.8	11.4 ± 14.5	0.174	11.8 ± 12.7	6.1 ± 8.8	0.008	9.4 ± 10.6	4.6 ± 7.9	0.006	6.0 ± 8.9	2.7 ± 4.6	0.022
Reflux symptom	14.5 ± 19.1	16.2 ± 15.4	0.218	16.8 ± 20.6	11.4 ± 15.0	0.262	14.2 ± 16.9	6.4 ± 11.2	0.002	13.7 ± 16.2	6.2 ± 10.8	0.007	10.3 ± 15.0	4.4 ± 6.7	0.084	8.8 ± 12.9	2.8 ± 5.3	0.001	5.1 ± 9.8	1.8 ± 4.7	0.033
Eating restrictions	8.8 ± 13.4	8.9 ± 11.5	0.651	5.9 ± 10.1	6.3 ± 10.9	0.878	6.3 ± 11.2	3.2 ± 6.9	0.094	5.8 ± 11.1	4.2 ± 9.2	0.487	2.9 ± 7.0	2.3 ± 5.7	0.772	2.6 ± 5.7	1.4 ± 4.4	0.188	1.4 ± 3.5	1.0 ± 3.7	0.605
Having a dry mouth	16.2 ± 23.4	19.6 ± 21.0	0.192	4.4 ± 11.4	7.8 ± 16.4	0.231	7.4 ± 15.1	5.4 ± 12.4	0.486	6.9 ± 13.6	8.3 ± 17.6	0.802	6.9 ± 15.8	6.9 ± 18.7	0.688	8.2 ± 15.8	1.7 ± 7.3	0.005	3.1 ± 9.8	1.4 ± 6.7	0.302
Taste	8.8 ± 17.9	5.9 ± 14.0	0.291	6.4 ± 14.4	7.8 ± 15.4	0.527	6.9 ± 17.8	5.4 ± 12.4	0.940	5.9 ± 15.2	8.3 ± 16.7	0.286	6.4 ± 13.2	5.9 ± 17.2	0.402	2.9 ± 9.5	0.6 ± 4.3	0.083	1.9 ± 7.7	0	0.096
Anxiety	17.8 ± 20.3	18.6 ± 19.7	0.650	6.7 ± 12.3	9.2 ± 15.7	0.309	5.9 ± 11.3	6.9 ± 14.3	0.946	10.0 ± 15.7	7.0 ± 14.5	0.230	9.3 ± 15.9	6.9 ± 13.8	0.281	5.7 ± 11.7	1.7 ± 4.9	0.017	3.1 ± 8.2	0.9 ± 3.8	0.092
Body image	5.9 ± 14.0	6.9 ± 15.8	0.794	2.0 ± 7.9	3.9 ± 10.8	0.228	3.4 ± 10.2	3.4 ± 10.2	1.000	4.9 ± 11.9	4.9 ± 13.2	0.834	3.4 ± 11.7	2.5 ± 8.8	0.737	1.8 ± 7.5	0	0.073	1.2 ± 6.4	0	0.177
Hair loss	5.4 ± 10.9	6.1 ± 12.5	0.738	2.2 ± 8.6	2.9 ± 9.9	0.737	3.7 ± 10.3	2.2 ± 8.6	0.269	0.74 ± 4.5	2.0 ± 6.8	0.15	2.5 ± 9.7	2.0 ± 9.3	0.525	5.3 ± 12.3	1.1 ± 6.0	0.021	2.5 ± 8.8	0.7 ± 4.8	0.209

B-I group: Billroth-I group; EORTC: European Organization for Research and Treatment of Cancer; QoL: Quality of life; R-Y group: Roux-en-Y group.

Discussion

B-I and R-Y anastomoses are traditional reconstructions after distal gastrectomy. However, QoL between B-I and R-Y anastomoses after distal gastrectomy are still inconsistent in different studies. Furthermore, long-term dynamically assessments to the postoperative QoL of patients with relatively advanced stage at multiple time points were rare. Therefore, the certainty of current evidence for QoL was low because of the limited number of studies, the disparity of design, and conduction in different studies. And future studies should include long-term QoL after different reconstruction methods.^[22] This study compared the impacts of two reconstructive methods on the long-term QoL after distal gastrectomy and found that the scores of the global health status of the R-Y group were higher than those of the B-I group with statistical differences, and the reflux and epigastric pain were milder in the R-Y group than those of the B-I group at the postoperative 1-, 3-, and 5-year time points.

Regarding survival, the survival of patients with gastric cancer was associated with several important factors, such as TNM stage, lymphadenectomy, curative degree, and adjuvant treatments.^[4–6] Theoretically, digestive tract reconstruction was not the independent prognostic factor for gastric cancer patients. Our results have proven that there were no significant differences in terms of overall survival and recurrence-free survival between the two groups. Nakamura et al^[8] also demonstrated that no significant difference was found between the two groups on the 3-year overall survival.

QoL has been regarded as an important outcome measurement parameter and emphasized in cancer patients, provided that there were no significant differences in overall survival rate and recurrence-free survival rate between the two groups. Several studies have compared the effects of different reconstructions after distal gastrectomy on postoperative QoL. Some studies have shown that R-Y anastomosis can reduce reflux symptoms and reflux-related gastritis and esophagitis, thus improving the QoL.^[12,23–26] Meta-analyses compared B-I, B-II, and R-Y anastomoses after distal gastrectomy, and showed a significantly lower incidence of reflux symptoms in the R-Y group than the other two groups, although R-Y reconstruction was considered as technical complexity.^[27,28] A study assessed the efficacy of R-Y and B-I reconstructions after laparoscopic-assisted distal gastrectomy and reported a significant increase in the incidence of heartburn 12 months after surgery in the B-I group.^[15] However, the studies mentioned above mainly evaluated the short-term QoL. For long-term QoL, Okuno et al^[29] have demonstrated that the B-I group had more frequent heartburn than the R-Y group at 3 and 5 years postoperatively. Additionally, the incidence of heartburn in patients with B-I anastomosis had gradually increased until their 5-year postoperative follow-up. Nunobe et al^[26] demonstrated that patients were less likely to experience heartburn after R-Y anastomosis than after B-I anastomosis. However, the data were retrospectively obtained with an unverified medical chart. The present RCT prospectively assessed long-term QoL at postoperative 36 and 60 months with the EORTC QLQ-C30 and stomach module QLQ-STO22 questionnaires, whose clinical practicability and validity have been proven in many studies,^{[20,21,28–30]} and found that the R-Y group was better in controlling the reflux symptom than that of the B-I group. Indeed, the endoscopic examination at postoperative 1 year found that the degree and extent of remnant gastritis were milder in the R-Y group,^[9] which supported the finding of QoL. However, inconsistent results have still existed. Another RCT study found no significant differences on postoperative QoL among B-I, B-II, and R-Y during the 12 months after distal gastrectomy.^[11] Nakamura et al^[8] also found that although the incidences of remnant gastritis were significantly lower following R-Y reconstruction at postoperative 12 and 36 months, the reflux symptoms were not significantly different. The results of meta-analyses also suggested that there may be little to no difference on QoL between R-Y and B-I reconstruction, although R-Y reconstruction probably reduced the incidence of bile reflux compared to B-I reconstruction.^[22]

Pain is a kind of subjective feeling, which is influenced by many factors, such as postoperative intraperitoneal adhesions,^[31] the incidence and severity of reflux symptom and cholelithiasis, etc. Several studies showed that no significant difference on epigastric pain was observed between the two kinds of anastomoses at postoperative 3 or 5 years.^{[8,13,22,26,32]} However, the present study found that R-Y anastomosis was associated with milder epigastric pain significantly at postoperative 1, 3, and 5 years. Du et al^[33] also reported that patients undergoing R-Y reconstruction had significantly less pain, compared with B-I reconstruction. The decreased reflux and reflux-related remnant gastritis in the R-Y group also could contribute to alleviating the epigastric pain symptom and subsequently improving the global health status at postoperative 1, 3, and 5 years.

Several reasons may partly explain the inconsistent results of QoL between different studies. Firstly, differences of study designs, patient selections, operative procedures, outcome measurements, and various questionnaires adopted to evaluate the QoL might lead to different outcomes. Secondly, QoL is a kind of subjective perception, which might prone to be affected by many factors.

No patient was lost to follow-up in this trial. In the follow-up work, first of all, we had special personnel to follow up patients regularly, and provided consultation and suggestions for patients during the follow-up process, which could increase the compliance of patients. Indeed, we followed up patients more frequently than the follow-up time points specified for this study. And several follow-up methods included hospital out-patients follow-up, telephone follow-up, online follow-up, letter follow-up, etc, were applied to reduce the lost follow-up. When patients were discharged from the hospital, we repeatedly advised patients to visit the hospital regularly for follow-up. At the time of admission, we collected the phone numbers and home addresses of at least two or more relatives of the patients to avoid the loss of follow-up due to changing mobile phone numbers or moving. We would also set up a WeChat group for patients, through which we could solve the difficulties encountered by patients and obtain information for follow-up consultations in other hospitals. All of these methods are beneficial in reducing the rate of loss of follow-up.

There are some limitations of this study. Firstly, the QoL by questionnaires was a kind of subjective data and prone to be prejudiced and biased by many aspects (eg, patients’ feeling, the competency of communication or comprehension, and cultural background). Although the results are less convincing when compared to the objective outcome measurements, this was inevitable and widely accepted in QoL-related studies. Secondly, this was just a single-center trial, which might limit the generalization of our results. Finally, other outcome parameters, such as nutritional status, food intake, or blood glucose, had not been evaluated in this trial. That is because these outcomes have been well addressed in other studies. Meanwhile, the improvements of these parameters would meliorate the QoL ultimately. Furthermore, the QoL could reflect the real status of patients. Notably, the QoL of patients might not be impaired, even though the reflux remnant gastritis was significantly more frequent in the B-I group. Therefore, QoL was designed as the primary and core outcome in the present trial.

In conclusion, there were no significant differences on survival between the B-I group and the R-Y group. R-Y reconstruction was associated with better long-term QoL by reducing reflux and epigastric pain, and promoted a better global health.

Acknowledgement

This work was internally supported by the Volunteer Team of Gastric Cancer Surgery (VOLTGA), West China Hospital, Sichuan University, China.

Funding

Domestic support from the Foundation of Science & Technology Department of Sichuan Province (Nos. 2020YJ0212 and 2023YFS0060) and 1. 3. 5 project for disciplines of excellence, West China Hospital, Sichuan University (No. ZYJC21006).

Conflicts of interest

None.