Traditional Chinese Medicine Enhances Survival in Patients with Gastric Cancer after Surgery and Adjuvant Chemotherapy in Taiwan: A Nationwide Matched Cohort Study

Background Gastric cancer remains a major cancer globally. More than half of patients with gastric cancer undergo surgery in Taiwan; however, few large nationwide studies have investigated the effects of traditional Chinese medicine (TCM) on gastric cancer management after surgery. This study aimed to evaluate the effect of TCM on patients with gastric cancer following surgery and adjuvant chemotherapy in Taiwan and its prescription trends. Methods and Materials The cohort sampling data set was obtained from the Registry of Catastrophic Illness Patient Database, a research database of patients with severe illnesses from the National Health Insurance Research Database, Taiwan. Patients who had received a new diagnosis of gastric cancer and had undergone surgery were enrolled. We matched TCM users and nonusers at a ratio of 1 : 3 based on the propensity score, and TCM users were also grouped into short-term and long-term users. Results The number of TCM users and nonusers was 1701 and 5103 after applying the propensity score at a ratio of 1 : 3. Short-term users and long-term TCM users were independently associated with a decreased risk of death with HRs of 0.59 (95% confidence interval (CI), 0.55–0.65) and 0.41 (95% CI, 0.36–0.47), respectively, compared with TCM nonusers. We also obtained similar results when we adjusted for covariates in the main model, as well as each of the additional listed covariates. We also observed similar HR trends in short-term users and long-term TCM users among men and women aged <65 years and ≥65 years. The most commonly prescribed single herb and herbal formula in our cohort were Hwang-Chyi (Radix Hedysari; 11.8%) and Xiang-Sha-Liu-Jun-Zi-Tang (15.5%), respectively. Conclusion TCM use was associated with higher survival in patients with gastric cancer after surgery and adjuvant chemotherapy. TCM could be used as a complementary and alternative therapy in patients with gastric cancer after surgery and adjuvant chemotherapy.


Introduction
Gastric cancer, including adenocarcinomas and other carcinomas, remains a major cancer globally and is ranked fifth among the most frequently diagnosed cancers, with over 1, 000, 000 new cases reported in 2018, and it is the third leading cause of cancer-related death, with approximately 7, 83, 000 deaths. [1] e incidence rate of gastric cancer is markedly high in Eastern Asia with 32.1 per 1, 00, 000 among men and 13.2 among women, and the incidence rate in South Korea is approximately 60 per 1, 00, 000 new cases annually among men and 25 per 1, 00, 000 among women. [2] Surgical resection is often adopted in the early stages of gastric cancer, whereas adjuvant therapies, such as radiotherapy or chemotherapy, are often considered in the advanced stages of gastric cancer. [3,4] In 2010-2014, the five-year survival rate of gastric cancer was 69% in South Korea, 60% in Japan, and less than 40% in most other countries, including Taiwan and the United States. [5].
Traditional Chinese medicine (TCM), traumatology manipulative therapies, and acupuncture, which are included in traditional Chinese medical in Taiwan, are widely accepted as some of the most popular complementary and alternative (CAM) therapies for patients with cancer. In addition, Chinese herbal products (CHPs) represent decoctions that have been adopted more recently based on their consistent quality and are applied in clinical settings. In addition, because of their quality, convenience, and approval for full reimbursement for both single herbs and herbal formulas CHPs by the National Health Insurance (NHI) in Taiwan, Chinese medical physicians prescribe them to their patients. A computerized reimbursement database of the NHI, the National Health Insurance Research Database (NHIRD), contains comprehensive data on both TCM and Western medicine. Consequently, the NHIRD offers an appropriate resource for conducting pharmacoepidemiological research on medicine use.
A previous study reported that TCM improved overall survival among patients with gastric cancer in Taiwan. [6] In Taiwan, most patients with gastric cancer undergo surgery, and more than half of the patients who undergo surgery receive chemotherapy afterward. ere are many clinical studies on the treatment of gastric cancer by TCM after surgery and chemoradiotherapy in China. However, large nationwide studies on the potential effects of the use of TCM on patients with gastric cancer after surgery and adjuvant chemotherapy remain limited. erefore, the aim of the present study was to evaluate the effect of the use of TCM on patients with gastric cancer following surgery and adjuvant chemotherapy in Taiwan and the associated TCM prescription trends.

Data Source.
e cohort sampling data set of the present study was obtained from the Registry of Catastrophic Illness Patient Database (RCIPD), a research database of patients with severe illnesses within the NHIRD system, NHI, Taiwan.
Because the Taiwanese NHI is obligatory for all residents, the NHIRD, which covers almost the entire population of 23.7 million people, is a rather detailed database in a healthrelated field.
Using data from the RCIPD, we can gather information based on both admissions and outpatient visits, which contains patient data such as admission and discharge dates, sex, age, diagnoses (made according to the International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes), and outpatient visits. In addition, data include prescription for patients, such as the prescribed medicine, initiation date, duration, dosage, and total expenditure.
Because the information in the database is anonymous, in line with personal electronic data protection regulations set by the National Health Research Institutes of Taiwan, NHIRD data can only be used for research purposes. e use of the database, in which information is anonymized and informed consent requirement is waived, was approved by the Ethics Review Board of Chang Gung Memorial Hospital, Chiayi Branch, Taiwan.

Study Subjects.
e study cohort was selected from admission and outpatient department records using diagnostic variables in the RCIPD. e ICD-9-CM codes 151 and codes beginning with 151 were used for the identification of gastric cancer. We included patients who had received a diagnosis of gastric cancer between January 1, 1999, and December 31, 2008, and survival was tracked until December 31, 2013. e endpoint of observation was death or end of follow-up. We excluded patients who had previously received a diagnosis of other cancer. In addition, we excluded patients who had not undergone surgery, patients who had undergone chemotherapy before surgery, and patients who did not undergo chemotherapy after surgery. In the consequence, patients with gastric cancer after surgery and adjuvant chemotherapy were included, whether they have done radiotherapy or not. Patients with missing information related to age, sex, level of urbanization, or level of income were also excluded.

Traditional Chinese Medicine
Exposure. We downloaded data on reimbursements for Chinese medicine from the NHI website. We obtained the corresponding drug data from the Committee on Chinese Medicine and Pharmacy website, including the proportions of each constituent, date and period of drug approval, drug names, and manufacturers' codes.
To classify patients into TCM users and nonusers, patients who used TCM for 30 days or more were considered TCM users and those who used TCM for less than 30 days were considered TCM nonusers. In addition, to enable the evaluation of a dose-response relationship, we further grouped TCM users into short-term TCM users (patients who used TCM for 30-179 days) and long-term TCM users (patients who used TCM for more than 180 days).

Matched
Cohort. By using propensity scores, TCM users and nonusers were perfectly matched at a ratio of 1 : 3, and it is a great estimate of how likely a patient would use TCM as a treatment. Background information, such as urbanization level, monthly insurance income, age, sex, and all comorbidities, was obtained.

Statistical Analysis.
We compared the distribution of baseline characteristics, namely sex, age at surgery, urbanization level, income level, comorbidities, and chemotherapy regimen, between TCM users and nonusers. Evidence-Based Complementary and Alternative Medicine We used the Kaplan-Meier method to estimate the cumulative probability of the overall survival of TCM users and nonusers. We performed the log-rank test to examine differences in overall survival between the cohorts. We used the Cox proportional hazards model to compute hazard ratios (HRs) at a 95% confidence interval (CI) after adjustment for age, sex, monthly insurance income, urbanization level, and chemotherapy regimen. A two-tailed P value of 0.05 was considered significant. All analyses were conducted using SAS v9.4 (SAS Institute, Cary, NC, USA).

Results
Of 33, 256 patients who had received a new diagnosis of gastric cancer between 1999 and 2008, 1331 were excluded because they had received a prior diagnosis of another cancer, 21 103 were excluded because they did not undergo surgery, chemotherapy, or surgery after chemotherapy, 2 were excluded because they were aged less than 18 years, and 203 were excluded because their data were incomplete. A total of 1721 and 8896 patients were classified as TCM users and nonusers, respectively. After applying the propensity score at a ratio of 1 : 3, the number of TCM users and nonusers was 1701 and 5103, respectively ( Figure 1).
In our study cohort, the mean age at surgery of TCM users and nonusers was 59.8 and 59.7 years, respectively. No significant differences were noted between the percentage of TCM users and nonusers based on sex, age at surgery, urbanization level, income level, and comorbidities (Table 1). Table 2 shows overall survival in patients with gastric cancer among TCM nonusers, short-term TCM users, and long-term TCM users. Figure 2 shows Kaplan-Meier curves for overall survival in patients with gastric cancer among TCM nonusers, short-term TCM users, and long-term TCM users.
After adjusting for sex, age, urbanization level, and income level in the main model, short-term and long-term TCM users were independently associated with a decreased risk of death with HRs of 0.59 (95% CI, 0.55-0.65) in TCM users and 0.41 (95% CI, 0.36-0.47) in TCM nonusers. We obtained similar results when covariates were adjusted for in the main model, as well as for each additional covariate, namely, diabetes mellitus, hypertension, alcoholism, smoking-related disorder, chronic renal failure, liver cirrhosis, and chemotherapy regimen. In the subgroup analysis of sex and age, we observed similar HR trends in short-term and long-term TCM users among men ( (Table 3). Table 4 lists single herbs and herbal formulas most commonly prescribed for gastric cancer after surgery and adjuvant chemotherapy in our cohort, including the prescription frequencies, average prescription durations, and average daily doses. e single herb most commonly prescribed in our cohort was Hwang-Chyi (Radix Hedysari; 11.8%), followed by Dan-Shen (Radix Salviae Miltiorrhizae; 9.8%), Yan-Hu-Suo (Rhizoma Corydalis; 9.4%), Bai-Hua-She-She-Cao (Herba Hedyotidis Diffusae; 9.2%), and Hou-Pu (Cortex Magnoliae; 9.0%). Xiang-Sha-Liu-Jun-Zi-Tang (15.5%) was the herbal formula most commonly prescribed in our cohort, followed by Ping-Wei-San (12.6%), Ban-Xia-Xie-Xin-Tang (11.8%), Bu-Zhong-Yi-Qi-Tang (10.3%), and Shen-Lin-Bai-Zhu-San (10.2%). e daily doses of the most commonly prescribed single herbs and herbal formulas were in the ranges of 1.4-1.7 g and 4.0-4.9 g, respectively. e average duration of single herbs and herbal formulas most commonly prescribed ranged from 7.8 to 11.1 days.

Discussion
is study has several merits. Because our population data set was directly obtained from the RCIPD, our study enrolled all patients with gastric cancer between 1999 and 2008 in Taiwan instead of a representative sample. In addition, our study cohort was obtained from the RCIPD, which is highly accurate due to its high demand for implicating catastrophic illness.
A previous study demonstrated that TCM enhanced the survival of patients with gastric cancer. [6] As the tumor stage could not be defined, the previous study could not  Evidence-Based Complementary and Alternative Medicine 3 reveal the potential relationship between adopted therapies and differences in disease severity. However, because in the present study we examined patients with gastric cancer who underwent chemotherapy following surgery, our study cohort had similar disease severity levels and stages. Based on propensity score matching, we increased homogeneity between TCM users and nonusers and reduced the number of nonmatched users as much as possible. After propensity score matching, only 20 (1.16%) of 1721 TCM users could not be matched in our study, whereas more than 25% of TCM users could not be matched in the previous study. In addition, we classified TCM users into short-term users who used TCM for less than 180 days and long-term users who used TCM for 180 days or more to observe the effect of long-term TCM use. To verify our results, we adjusted additional covariates, namely, diabetes mellitus, hypertension, alcoholism, smoking-related disorder, chronic renal failure, liver cirrhosis, and chemotherapy regimen, to examine whether results were consistent, and similar results in subgroup effects based on sex and age at surgery also confirmed our findings. Finally, because survival was calculated from the day TCM was taken in our study, we reduced immortal time bias as much as possible.
In Taiwan, CHPs are considered a modern form of decoctions because of their more consistent quality when compared with TCM. e NHI program reimburses patients fully for both single-herb products and multi-herb products prescribed. Chinese medical physicians used to prescribe   Evidence-Based Complementary and Alternative Medicine one or more herbal formulas in combination with several single herbs in each prescription according to the state of illness of a patient. In our study, half of the single herbs and herbal formulas most commonly prescribed that are listed in Table 4 are often used to treat ailments of the gastrointestinal system. Xiang-Sha-Liu-Jun-Zi-Tang, the most commonly prescribed herbal formula identified in the present study, was often used to treat upper gastrointestinal disorders, including indigestion, gastroesophageal reflux, and anorexia, [7,8] and has been reported to prevent nausea, vomiting, and anorexia induced by chemotherapy in patients with cancer [9][10][11] and exert beneficial effects on gastrointestinal disorders, anorexia, and postoperative gastric ileus when used in  Evidence-Based Complementary and Alternative Medicine combination with Western drugs. [12,13] Ping-Wei-San is also often prescribed for the treatment of upper gastrointestinal disorders, including esophageal reflux, gastritis, gastric or duodenal ulcers, and enteritis. [14][15][16] Furthermore, Ban-Xia-Xie-Xin-Tang has been reported to reduce diarrhea during chemotherapy [17,18] and could prevent chemotherapy-induced oral mucositis in esophageal cancer [19] and in chemoradiation of head and neck cancers. [20] Shen-Ling-Bai-Zhu-San is one of the most common formulas used for the treatment of ulcerative colitis [21,22] and has also been reported to reduce gastric cancer chemotherapy-induced toxicity. [23] Hou-Pu (Cortex Magnoliae) had been demonstrated to be beneficial for the treatment of gastrointestinal disorders. [24] Huang-Chyi (Radix Hedysari), which was the most commonly prescribed single herb in our study, and Bu-Zhong-Yi-Qi-Tang have been demonstrated to alleviate cancer-related fatigue. [25][26][27] Dan-Shen (Radix Salviae Miltiorrhizae) and Bai-Hua-She-She-Cao (Herba Hedyotidis Diffusae) have been shown to have anticancer effects, [28][29][30] whereas Bai-Hua-She-She-Cao (Herba Hedyotidis Diffusae) is the most common single herb prescribed for patients with colon cancer and breast cancer in Taiwan. [31,32] Finally, Yan-Hu-Suo (Rhizoma Corydalis) showed pain control benefits. [33,34]. Our study has some limitations that should be addressed. First, the NHI program only reimburses the cost of CHPs prescribed by Chinese medical physicians. erefore, CHPs or decoctions purchased directly from TCM pharmacies were not included in our analysis. Consequently, the frequency of TCM use could have been underestimated in the present study. Nevertheless, because the NHI program has a comprehensive cover for CHPs prescribed by Chinese medical physicians, which is generally lower than the cost of CHPs or decoctions sold in TCM pharmacies, the underestimation would likely be minimal. In the present study, we could not verify exact dosages taken by enrolled patients. We could only assume that patients took medications as prescribed, which may overestimate the actual ingested dosage because some degree of noncompliance is always anticipated. Furthermore, detailed information, such as living environment, lifestyle, nutrition, and other examination data, were not available from the RCIPD. Finally, the stages of cancer were not recorded in the RCIPD. Nevertheless, we controlled the severity of cancer in our study cohort to some extent by selecting patients with gastric cancer undergoing chemotherapy after surgery.

Conclusions
In conclusion, in our study, TCM use was associated with higher survival in patients with gastric cancer after surgery and adjuvant chemotherapy. TCM could be adopted as a CAM therapy in patients with gastric cancer after surgery and adjuvant chemotherapy.

TCM:
Traditional Chinese medicine CAM: Complementary and alternative medicine CHPs: Chinese herbal products NHI: National Health Insurance NHIRD: National Health Insurance Research Database RCIPD: Registry of Catastrophic Illness Patient Database ICD-9-CM: International Classification of Diseases, Ninth Revision, Clinical Modification HRs: Hazard ratios CI: Confidence interval.
Data Availability e data sets used and analyzed in this study are available from the corresponding author on reasonable request. Disclosure e funding body played no role in study design, analysis, and interpretation of data in this paper. Also, the statistical results and conclusions presented in this paper do not represent those of the Bureau of National Health Insurance, Department of Health, or the National Health Insurance Research Institutes.

Conflicts of Interest
e authors declare that they have no conflicts of interest.