Patient characteristics according to MA prescription status
We screened 13,358 adult patients with newly diagnosed metastatic digestive malignancies presenting during the index period. Among these patients, we excluded 72 patients with a history of MA prescription at baseline, 6,038 patients who were treated with curative surgery and CCRT, 933 patients who survived for more than two years following the index date without any active anticancer therapy, and 395 patients with double primary cancer that occurred during the follow-up period. Then, 3,982 patients with non-gastric cancer were excluded. Finally, a total of 1,938 patients with metastatic gastric cancer were included in the current analysis (Fig. 1).
Baseline characteristics according to MA prescription patterns are listed in Table 1. There were 1,260 patients in the MA prescription group (out of a total of 1,938; 65.0%). We found that younger patients with a lower mCCI as well as those treated in a tertiary hospital were more likely to be prescribed MA. Among patients treated or not treated with palliative chemotherapy, MA was prescribed in 974/1,251 (77.9%) and 286/687 (41.6%) patients, respectively. On multivariate analysis, we found that elderly patients were more frequently prescribed MA than younger patients (OR [odds ratio] for a 10-year increase, 1.138, 95% CI 1.049-1.234, p=.002; Online Resource 3). Palliative chemotherapy was the most important factor predicting MA prescription (OR 5.612, 95% CI 4.413-7.137, p<.001). The proportion of patients treated with palliative chemotherapy decreased with increasing age (from 94.2% in patients aged <40 years to 23.3% in those aged ≥80 years). This indicates that a higher rate of MA prescription in younger patients on univariate analysis was resulted from a higher proportion of receiving palliative chemotherapy in those patients.
Prescription patterns for MA
The continuity of MA prescription was assessed in 2,921 patients who were prescribed MA at least twice (Online Resource 4). Continuous and intermittent prescription was performed in 343 (37.0%) and 583 (63.0%) patients, respectively. Elderly patients and patients not treated with palliative chemotherapy were more commonly found in the continuous prescription group (vs. the intermittent prescription group).
The median time from the index date to the first MA prescription was 31 days (IQR [interquartile range], 8-92 days). The median time became shorter with increasing age (59 days in those aged <40 years, 40 days in those aged 40-49 years, 37 days in those aged 50-59 years, 36 days in those aged 60-69 years, 28 days in those aged 70-79 years, and 10 days in those aged ≥80 years). The median total number of MA prescription days was 39 days (IQR 16-83). The most frequent range for the average daily MA dose was 600-<800 mg/day (in 43.0% of evaluated patients), followed by 800-<1,000 mg/day (28.4%) and 400-600 mg/day (20.2%). Average daily MA doses of <400 mg/day and ≥1,000 mg/day were observed in 2.4% and 6.0% of the evaluated patients, respectively.
Survival
Differences in survival according to the MA prescription patterns were assessed according to whether palliative chemotherapy was administered. In patients treated with palliative chemotherapy, the prescription of MA did not influence survival outcome (Fig. 2A). However, in those who did not receive palliative chemotherapy, the group prescribed MA exhibited a longer survival time compared to the group that was not prescribed MA (Fig. 2B).
However, in multivariate analyses, we found that MA prescription had little impact on survival. Although there was a statistically significant difference in survival, the HR for MA prescription was only 1.001 in patients treated with palliative chemotherapy. There was no statistically significant difference in survival between the MA and non-MA prescription groups in patients not treated with palliative chemotherapy (Online Resource 5).
Venous thromboembolism
VTE was newly diagnosed in 55/1,427 (3.9%) patients with metastatic gastric cancer during the follow-up period (Table 2). The incidence of VTE was higher in the MA prescription group than in the non-prescription group (4.3% vs. 2.9%).
However, on multivariate analysis (Table 3), MA prescription was not a predictive factor for VTE occurrence (OR 1.310, 95% CI 0.692-2.480, p=0.41). When the analysis was stratified by treatment status, we found that there were no statistically significant differences in VTE occurrence between MA prescription and non-prescription groups (Online Resource 6). When considering prescription patterns for MA (Online Resource 7), we found that neither continuous prescription nor the average daily dose of MA had any impact on the incidence of VTE.