We performed a retrospective cohort study using the diagnostic procedure combination (DPC) databases of nine hospitals between January 2014 and March 2019. The combined database comprised the records of all inpatients and outpatients at the University of Tokyo Hospital, Shuto General Hospital, Fukui Prefectural Hospital, Nerima Hikarigaoka Hospital, St. Luke’s International Hospital, Toyonaka Municipal Hospital, Ishikawa Prefectural Central Hospital, and Nagasaki Minato Medical Center and of inpatients at Tonan Hospital. The database included diagnoses, comorbidities, and adverse events using the International Classification of Diseases, 10th revision and Japanese original disease codes. It also included the cancer stage according to the Union for International Cancer Control classification system[9], Japanese original medication and procedure codes, and Barthel index (BI)[10].

From the database, we identified patients who had undergone palliative surgery or SEMS placement for gastrointestinal obstruction, including esophageal bypass surgery, gastrojejunostomy, duodenojejunostomy, intestinal bypass surgery, stoma surgery, esophageal stenting, gastroduodenal stenting, or colonic stenting, who did not undergo gastrointestinal tract resection thereafter. We excluded patients without malignant disease, those with a history of gastrointestinal perforation or fistula, and patients with multifocal gastrointestinal obstructions. The codes used for patient selection are listed in Supplementary Table 1.

We selected the chemotherapy group (patients who received any chemotherapy drugs after the intervention) with the BSC group (patients who did not receive chemotherapy drugs after the intervention) (Figure 1). The follow-up period was from the date of the intervention to death or the final visit. The end of follow-up was March 2019, and loss to follow-up was defined as the date of the final visit. The study was approved by the Institutional Review Board of the University of Tokyo Hospital (No. 2019161NI).

The primary outcome was overall survival. The secondary outcomes were patency duration and adverse events, including perforation and gastrointestinal bleeding. Patency duration was defined as the time between the first food intake after the intervention and reintervention, stopping food intake, or death. Perforation was defined as surgery for suture, drainage, or intra-abdominal lavage. Gastrointestinal bleeding was defined as any gastrointestinal bleeding requiring endoscopic hemostasis. The procedure codes for outcomes are listed in Supplementary Table 2.

We evaluated the following clinical factors: Age, sex, comorbidities, BI, medication use, cancer type, cancer stage, obstruction site, chemotherapy before the intervention, and intervention type. Age was categorized into < 75 years and ≥ 75 years. Comorbidities were evaluated by the Charlson comorbidity index (CCI)[11] and categorized as < 3 and ≥ 3. BI was categorized as ≥ 60, < 60, and missing. We evaluated the use of aspirin, thienopyridine, warfarin, direct oral anticoagulants (including dabigatran, rivaroxaban, apixaban, and edoxaban), other antiplatelet drugs (including dilazep hydrochloride hydrate, dipyridamole, trapidil, cilostazol, limaprost alfadex, ethyl icosapentate, beraprost sodium, sarpogrelate hydrochloride, and ozagrel sodium), nonsteroidal anti-inflammatory drugs, and steroids. The cancer type was categorized into esophageal cancer, gastric cancer, pancreatic cancer, colorectal cancer, and other cancers. The cancer stage was categorized into stage I–III, stage IV or recurrence, and missing. The obstruction site was categorized as esophageal obstruction, gastroduodenal obstruction, or lower gastrointestinal obstruction. The intervention type was categorized as palliative surgery or SEMS placement. The International Classification of Diseases, 10th revision codes of primary cancers and comorbidities are listed in Supplementary Table 3, and the medication codes are shown in Supplementary Table 4.

Overall survival and patency durations were estimated by the Kaplan–Meier method and were compared by log-rank test. Data were censored at the date of the final visit. Univariate Cox proportional hazards models were used to estimate crude hazard ratios and 95% confidence intervals (CIs). The multivariate Cox proportional hazards models were used to estimate adjusted hazard ratios (aHRs) using age, sex, cancer type, cancer stage, CCI, BI, and intervention type.

Categorical data were compared by the chi-squared test or Fisher’s exact test and continuous data by the Wilcoxon rank-sum test. A P value < 0.05 was considered indicative of statistical significance. Statistical analysis was performed using SAS software v. 9.4 (SAS Institute, Cary, NC, United States).

Fourteen hundred forty patients who had undergone palliative surgery or SEMS placement for unresectable gastrointestinal obstruction were extracted from the DPC database. After excluding patients without malignant disease (n = 234), those with a history of gastrointestinal perforation or fistula (n = 74), and patients with multifocal gastrointestinal obstructions (n = 10), the remaining 1122 patients were analyzed. In total, 470 patients who received chemotherapy drugs after the intervention (chemotherapy group) and 652 patients who did not receive chemotherapy drugs after the intervention (BSC group) were analyzed (Figure 2). The patients’ baseline characteristics are listed in Table 1. The age, CCI, BI, medication, cancer type, and cancer stage distributions were significantly different between the groups. The chemotherapy group had higher rates of < 75 aged patients, CCI ≥ 3, BI ≥ 60, and stage IV or recurrence.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Flow chart of patient selection. SEMS: Self-expandable metal stent.

During the follow-up period of 54.1 mo, the median overall survival durations were 19.3 mo (95%CI: 16.2-25.9 mo) in the chemotherapy group and 5.4 mo (95%CI: 3.6-7.6 mo) in the BSC group (log-rank test, P < 0.01; Figure 2). The median patency durations were 9.7 mo (95%CI: 7.7-11.5 mo) in the chemotherapy group and 2.5 mo (95%CI: 2.0-2.9 mo) in the BSC group (log-rank test, P < 0.01; Figure 3).

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Kaplan–Meier estimates of overall survival and patency duration of the chemotherapy group and the best supportive care group. A: Overall survival; B: Patency duration. BSC: Best supportive care.

The factors affecting overall survival are shown in Table 2. Multivariate analysis showed that the factors affecting overall survival were chemotherapy after the intervention (aHR, 0.36), CCI ≥ 3 (aHR, 1.56), BI < 60 (aHR, 2.04), gastric cancer compared with esophageal cancer (aHR, 0.64), colorectal cancer compared with esophageal cancer (aHR, 0.47), stage IV or recurrence compared with stage I–III (aHR, 1.79), chemotherapy before the intervention (aHR, 1.66), and SEMS placement compared with palliative surgery (aHR, 1.63).

The factors affecting patency are shown in Table 3. Multivariate analysis showed that the factors affecting patency duration were chemotherapy after the intervention (aHR, 0.49), CCI ≥ 3 (aHR, 1.41), BI < 60 (aHR, 1.55), colorectal cancer vs esophageal cancer (aHR, 0.67), stage IV or recurrence compared with stage I–III (aHR, 1.65), chemotherapy before the intervention (aHR, 1.64), and SEMS placement compared with palliative surgery (aHR, 2.48).

The results of a subgroup analysis of adjusted HRs for overall survival and patency duration were consistent with those of the overall analysis (Table 4).

The rates of adverse events are listed in Table 5. The perforation rate was 1.3% (6/470) in the chemotherapy group (3 gastric cancers, one colorectal cancer, 1 breast cancer, and 1 unclassifiable cancer) and 0.9% (6/652) in the BSC group (3 colorectal cancers, 2 gastric cancers, and 1 esophageal cancer) (P = 0.567). In 4 of the 6 perforation cases in the chemotherapy group, perforation occurred a mean of 137 d after chemotherapy initiation.

The gastrointestinal bleeding rate was 1.5% (7/470) in the chemotherapy group (4 gastric cancers, 1 esophageal cancer, 1 pancreatic cancer, and o1colorectal cancer) and 0.5% (3/652) in the BSC group (1 esophageal cancer and 2 pancreatic cancers) (P = 0.105). In 4 of 7 bleeding cases in the chemotherapy group, gastrointestinal bleeding occurred a mean of 294 d after chemotherapy initiation.

Malignant gastrointestinal obstruction is an important issue in advanced cancer and occurs in approximately 30% of patients with gastrointestinal cancer. Gastrointestinal obstruction causes oral intake impairment, nausea, vomiting, and abdominal pain and increases the risk of gastrointestinal perforation. Primary therapy involves fasting and decompression, and subsequently complete surgical resection is performed for resectable malignant gastrointestinal obstruction; palliative surgery includes bypass and stoma surgery or self-expandable metal stent (SEMS) placement.

The impacts of chemotherapy on patients with malignant gastrointestinal obstructions remain unclear, and multicenter evidence is lacking.

We performed a large multicenter cohort study to evaluate the effectiveness and safety of chemotherapy after palliative surgery or SEMS placement compared with best supportive care (BSC) in patients with unresectable malignant gastrointestinal obstructions. In addition, we aimed to identify the optimal population for chemotherapy after palliative surgery or SEMS placement.

We conducted a multicenter retrospective cohort study that compared the chemotherapy group who received any chemotherapeutics after interventions, including palliative surgery or self-expandable metal stent placement, for unresectable malignant gastrointestinal obstruction vs BSC group between 2014 and 2019 in nine hospitals. The primary outcome was overall survival, and the secondary outcomes were patency duration and adverse events, including gastrointestinal perforation and gastrointestinal bleeding.

In total, 470 patients in the chemotherapy group and 652 patients in the BSC group were analyzed. During the follow-up period of 54.1 mo, the median overall survival durations were 19.3 mo in the chemotherapy group and 5.4 mo in the BSC group (log-rank test, P < 0.01). The median patency durations were 9.7 mo [95% confidence interval (CI): 7.7-11.5 mo] in the chemotherapy group and 2.5 mo (95%CI: 2.0-2.9 mo) in the BSC group (log-rank test, P < 0.01). The perforation rate was 1.3% (6/470) in the chemotherapy group and 0.9% (6/652) in the BSC group (P = 0.567). The gastrointestinal bleeding rate was 1.5% (7/470) in the chemotherapy group and 0.5% (3/652) in the BSC group (P = 0.105).

Chemotherapy after interventions for unresectable malignant gastrointestinal obstruction was associated with increased overall survival and patency duration.

Our results showed that chemotherapy may be more beneficial in cases of pancreatic cancer and gastroduodenal obstruction. These findings will help guide future research on treatment approaches and precision medicine. In the near future, these precision medicine methods are expected to contribute to cancer therapies including molecular targeted anti-cancer drugs, monoclonal antibody therapy, and antibiotic therapies.