Traditional Chinese Medicine Enhances Survival in Cancer Patients on Immunotherapy or Chemotherapy

Abstract

Integrating Traditional Chinese Medicine (TCM) with Western medical treatments enhances cancer patients’ survival rates and quality of life. However, there is a lack of large-scale studies on the integration of TCM with immune checkpoint inhibitors (ICIs) or chemotherapy in cancer treatment. This retrospective cohort study analyzed data from the Taiwan National Health Insurance Research database and Taiwan Cancer Registry, including 2,512 patients treated with ICIs and 130,010 with chemotherapy between January 1, 2019, and June 30, 2021. Patients receiving TCM were classified as TCM users, while those who did not were nonusers. The primary outcome was all-cause mortality, assessed using Cox proportional hazards regression to estimate hazard ratios (HR) with 95% confidence intervals (CI), adjusting for demographics and comorbidities. Among ICI-treated patients, TCM users (n=720) had lower mortality rates than nonusers (n=1,792), with an incidence of 45.84 vs. 83.72 per 100 person-years (adjusted HR: 0.58; 95% CI: 0.51–0.66). Similarly, in chemotherapy-treated patients, TCM users (n=51,292) had reduced mortality (11.54 vs. 24.86 per 100 person-years; adjusted HR: 0.59; 95% CI: 0.57–0.60). These associations remained consistent across subgroup and sensitivity analyses. Commonly used TCM herbs included Hedyotis diffusa Willd (Bai Hua She She Cao), Scutellaria barbata (Ban Zhi Lian), and Astragali Radix (Huang Qi). These findings suggest that TCM use alongside ICIs or chemotherapy is associated with lower mortality, even after adjusting for confounders. Further prospective studies are needed to confirm these results and explore underlying mechanisms, advancing integrative cancer care.

Introduction

Cancer represents a significant global health burden and a major cause of death, both in Taiwan and worldwide. ¹ In 2022, approximately 20 million new cancer cases were reported globally, with approximately 9.7 million cancer-related deaths, accounting for 16.8% of all deaths and 22.8% of deaths attributed to noncommunicable diseases. ² The inherently complex and heterogeneous nature of cancer, combined with its high mortality rate, emphasizes the ongoing development of novel therapeutic strategies, including immunotherapy, targeted therapy, gene therapy, stem cell therapy, photodynamic therapy, nanoparticle-based treatments, and precision medicine. ³ Among these innovations, immune checkpoint inhibitors (ICIs) have emerged as a transformative treatment option.^4,5 ICIs function by targeting and inhibiting immune checkpoints such as PD-1, PD-L1, and CTLA-4, which regulate the immune response.^4,5 These checkpoints generally act as safeguards to prevent excessive immune activation and tissue damage. Nevertheless, cancer cells exploit these mechanisms to evade immune detection and destruction. ⁶ By blocking these checkpoints, ICIs restore immune system activity against cancer cells. Recent clinical trials have demonstrated the efficacy of ICIs across various cancer types, reshaping the treatment landscape for advanced malignancies. ⁵

Traditional Chinese Medicine (TCM) has a long history in cancer management, with the concept of “tumor” recorded in oracle bone inscriptions more than 3500 years ago. TCM adopts a holistic approach, emphasizing the interconnection of the body’s systems and the balance of Yin and Yang, as well as syndrome differentiation. These principles have guided the understanding of cancer etiology and treatment strategies. Today, TCM is widely recognized as a complementary modality in cancer care, with studies demonstrating its anticancer properties through mechanisms such as tumor suppressor gene reactivation, angiogenesis inhibition, modulation of tumor cell autophagy and apoptosis, mitochondrial regulation, DNA demethylation, histone acetylation, and tumor microenvironment modulation.^7,8 Furthermore, TCM exerts antimetastatic effects by inhibiting epithelial–mesenchymal transition processes. ⁹ From a TCM perspective, cancer treatment focuses on restoring immune balance, correcting metabolic dysfunction, and promoting phlegm elimination and anti-inflammatory activity. ¹⁰ Importantly, TCM improves the quality of life and mitigates depression in patients with cancer. ¹¹

Despite these advancements, the integration of TCM with novel cancer therapies, such as ICIs, remains underexplored. TCM treatments are inherently personalized, with therapeutic regimens tailored to individual patient conditions, and often involve complex herbal formulas rather than single-ingredient interventions. ¹² These characteristics pose challenges for conducting randomized controlled trials, making real-world studies a more suitable approach for evaluating the clinical role of TCM. ¹³ Evidence suggests that combining TCM with Western medical treatments improves survival outcomes and quality of life of patients.^14-17 Nevertheless, there are limited studies investigating the combined use of TCM and ICIs. ¹³ Therefore, the present study was conducted to investigate the impact of TCM on survival outcomes among patients receiving ICIs or chemotherapy. Moreover, the most frequently used TCMs among these patients were identified, providing insights into the potential benefits of integrating TCM into contemporary cancer care.

Methods

Data Sources

Data were obtained from Taiwan’s National Health Insurance Research Database (NHIRD) and the Taiwan Cancer Registry (TCR). The NHI program, introduced in 1995, is a compulsory health insurance plan with a coverage rate >99% of Taiwan’s residents, providing extensive reimbursements for healthcare expenses. The NHIRD has been widely used in epidemiologic research and has demonstrated good internal validity and data quality in previous studies. ¹⁸ The TCR includes data on cancer staging, primary site, diagnosis, treatment, and hospital follow-up information and can thus serve as a reference for evaluating cancer diagnosis and treatment data in the NHIRD. Cancer types were defined using ICD-O-3 codes, with detailed code definitions provided in Supplementary Table 1. All personal identifiers in the related databases have been encrypted and anonymized to ensure privacy protection. Hence, patient informed consent was not required. This study was approved by the Institutional Review Board of the Tri-Service General Hospital (IRB No. C202305093).

Study Design and Patient Selection

Using a cohort design, 253,094 patients with cancer who received antineoplastic medications between January 1, 2019, and June 30, 2021, were initially identified. Patients were excluded if they were aged <18 years, had previously received TCM outpatient services before initiating anticancer therapy, had an unspecified gender, or had received antineoplastic treatments outside the categories of ICIs or chemotherapy. Then, 132,522 eligible patients were subsequently categorized into two groups based on the Anatomical Therapeutic Chemical classification codes. The ICI group (n = 2,512) included patients treated with PD-1 and PD-L1 inhibitors (L01FF), because these are the only ICIs reimbursed by Taiwan’s NHI. The chemotherapy group (n = 130,010) comprised patients receiving alkylating agents (L01A), antimetabolites (L01B), plant alkaloids (L01C), cytotoxic antibiotics (L01D), and other antineoplastic agents (L01X). The index date was defined as the date of the first ICI treatment for patients in the ICI group and as the date of the first chemotherapy treatment for those in the chemotherapy group.

In Taiwan, many cancer patients use TCM treatments, supported by NHI funding for integrated care during hospitalization and continued access at outpatient clinics. In TCM, treatment procedures are not predefined, and each patient’s condition is addressed on an individual basis. To comprehensively evaluate the impact of TCM, each group was further stratified into subgroups based on TCM usage (720 TCM users and 1,792 non-TCM users in the ICI group; 51,292 TCM users and 78,718 non-TCM users in the chemotherapy group). Figure 1 depicts the sample selection process. To further characterize the clinical patterns of TCM use, the most frequently prescribed herbs and formulas were categorized into therapeutic functional groups based on their primary therapeutic actions, as defined by the WHO International Standard Terminologies on Traditional Medicine ¹⁹ and the Taiwan Herbal Pharmacopeia. ²⁰

Figure 1.

Patient selection steps

Study Outcomes

This study examined the association between TCM use and mortality among patients receiving ICIs or chemotherapy. The primary outcome was all-cause mortality, confirmed through cross-referencing with the national death registry. Patients were followed up from their index date until the occurrence of the primary outcome or December 31, 2021 (the final date in the study database), whichever occurred first.

Statistical Analysis

Descriptive statistics, including counts, percentages, mean values, and standard deviations, were used to summarize baseline characteristics, such as age, sex, cancer type, and comorbidities. To control for potential confounding, multivariable Cox proportional hazards regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs), adjusting for age, sex, personal income, cancer type, cancer stage, calendar year, and comorbidities (including hypertension, diabetes mellitus, hyperlipidemia, heart failure, stroke, chronic obstructive pulmonary disease, chronic kidney disease, depression, anxiety, and anemia), with detailed definitions provided in Supplementary Table 1. Interaction terms between TCM use and selected covariates were included in the Cox models to examine potential interactions. Additionally, a trend test was performed to evaluate the dose–response relationship between TCM utilization frequency and mortality risk. As sensitivity analyses, a 90-day landmark analysis was conducted, including only patients who survived at least 90 days after the index date, with follow-up initiated at the landmark time point. To further assess potential residual confounding, a negative control outcome analysis was performed using acute appendicitis (ICD-10-CM: K35), an outcome not expected to be associated with TCM use. All statistical analyses were conducted using the SAS software (version 9.4), with two-sided p < 0.05 considered statistically significant.

Results

Table 1 shows the baseline characteristics and comorbidities of patients receiving ICIs or chemotherapy, with or without concurrent TCM. Remarkably, 28.66% (720 of 2,512) of patients receiving ICIs used TCM, whereas 39.45% (51,292 of 130,010) of patients receiving chemotherapy used TCM. Regarding clinical stages, the majority of patients receiving ICIs were in stages 3 and 4, whereas those receiving chemotherapy had a more even distribution across stages 1 through 4. Patients receiving ICIs predominantly had liver, lung, or oral cavity and pharynx cancers, whereas those receiving chemotherapy had more varied types of cancer. The average annual number of TCM visits in the ICI and chemotherapy groups was approximately 8.25 and 13.27, respectively. All these factors were considered in the regression models to eliminate potential confounding effects.

Table 1.

Baseline Characteristics and Comorbidities of Cancer Patients Receiving Immune Checkpoint Inhibitor or Chemotherapy, With or Without Concurrent Traditional Chinese Medicine Treatment (N =132 522)

Variables	ICI therapy				Chemotherapy
	TCM users (n=720)		Nonusers (n=1,792)		TCM users (n=51,292)		Nonusers (n=78,718)
	n	%	n	%	n	%	n	%
Age, years (mean, SD)	62.66	11.32	65.09	11.83	59.39	13.83	63.81	13.78
Age Group
≤ 40	27	3.75	38	2.12	5130	10	4128	5.24
41∼65	384	53.33	887	49.5	28 189	54.96	38 282	48.63
≥ 65	309	42.92	867	48.38	17 973	35.04	36 308	46.12
Sex
Male	509	70.69	1419	79.19	21 636	42.18	47 637	60.52
Female	211	29.31	373	20.81	29 656	57.82	31 081	39.48
Index Year
2019	318	44.17	708	39.51	30 799	60.05	42 920	54.52
2020	296	41.11	743	41.46	14 864	28.98	24 111	30.63
2021	106	14.72	341	19.03	5629	10.97	11 687	14.85
Income, (TWD/month)
0∼18000	178	24.72	562	31.41	11 363	22.19	22 597	28.74
18000∼35000	365	50.69	873	48.8	24 973	48.76	37 550	47.76
>35000	177	24.58	354	19.79	14 883	29.06	18 480	23.5
Residential Area
Eastern District	18	2.5	41	2.29	1188	2.32	1980	2.52
Southern District	242	33.61	539	30.13	15 865	30.98	25 140	31.98
Central District	184	25.56	310	17.33	11 165	21.8	11 999	15.26
Northern District	276	38.33	899	50.25	22 999	44.91	39 504	50.24
Comorbidities
Diabetes Mellitus	163	22.64	457	25.5	10 049	19.59	19 338	24.57
Hypertension	271	37.64	722	40.29	17 238	33.61	31 408	39.9
Dyslipidemia	145	20.14	355	19.81	12 060	23.51	18 519	23.53
Comorbidities
Heart failure	10	1.39	49	2.73	1036	2.02	2562	3.25
Stroke	23	3.19	67	3.74	1327	2.59	2895	3.68
COPD	71	9.86	201	11.22	2121	4.14	5081	6.45
CKD	66	9.17	154	8.59	3592	7	7080	8.99
Depression	40	5.56	60	3.35	2550	4.97	3053	3.88
Anxiety	71	9.86	109	6.08	5723	11.16	5742	7.29
Anemia	22	3.06	38	2.12	1707	3.33	2390	3.04
Clinical stage at diagnosis
0	11	2.35	17	1.34	1326	4.16	1833	3.61
1	68	14.53	148	11.67	8026	25.17	10 067	19.80
2	70	14.96	168	13.25	8316	26.08	10 767	21.18
3	110	23.50	292	23.03	7848	24.62	13 296	26.15
4	209	44.66	643	50.71	6365	19.96	14 873	29.26
Cancer type at diagnosis
Liver	109	17.61	219	13.95	2387	5.96	4816	7.82
Lung	187	30.21	488	31.08	3713	9.27	7101	11.53
Oral Cavity and Pharynx	98	15.83	360	22.93	4059	10.13	8059	13.08
Colon and rectal	5	0.81	19	1.21	6065	15.13	10 745	17.44
Stomach	42	6.79	118	7.52	1588	3.96	2978	4.83
Bladder	56	9.05	134	8.54	2770	6.91	4338	7.04
Esophagus	6	0.97	10	0.64	1160	2.89	3126	5.07
Skin	28	4.52	60	3.82	73	0.18	207	0.34
Others	88	14.22	162	10.32	18 260	45.56	20 228	32.84
Number of visits to TCM (mean, SD)	8.25	13.52	-	-	13.27	20.53	-	-

Notes. ICI, immune checkpoint inhibitor; TWD, New Taiwan dollar; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; TCM, Traditional Chinese Medicine; SD, standard deviation.

Table 2 shows the comparison of mortality risk between TCM and non-TCM users in the ICI or chemotherapy population. In ICI-treated patients, the mortality rates per 100 person-years for TCM and non-TCM users were 45.84 (95% CI: 41.40–50.62) and 83.72 (95% CI: 79.12–88.50), respectively, whereas in chemotherapy-administered patients, the respective rates were 11.54 (95% CI: 11.33–11.76) and 24.86 (95% CI: 24.58–25.15). The Cox proportional hazards model showed that TCM users had lower mortality rates than non-TCM users among patients receiving ICIs or chemotherapy. In the ICI population, the crude HR for TCM users was 0.57 (95% CI: 0.51–0.64), and the adjusted HR was 0.58 (95% CI: 0.51–0.66), indicating a reduced risk compared with that in non-TCM users. This effect was even more pronounced in the chemotherapy group, where the crude HR was 0.48 (95% CI: 0.47–0.49), and the adjusted HR was 0.59 (95% CI: 0.57–0.6). In sensitivity analyses, the 90-day landmark analysis showed a similar protective association as the primary analysis, with adjusted HRs of 0.91 (0.76–1.09) in the ICI group and 0.89 (0.86–0.92) in the chemotherapy group. In addition, the negative control outcome analysis using acute appendicitis showed no significant association with TCM use (chemotherapy group: adjusted HR: 0.92, 95% CI: 0.75–1.15; ICI group: adjusted HR: 0.89, 95% CI: 0.06–12.92), suggesting that residual confounding is unlikely to fully explain the observed findings.

Table 2.

Comparison of Mortality Rates between Traditional Chinese Medicine Users and Non-Users Receiving Immune Checkpoint Inhibitor or Chemotherapy Treatments

Outcomes	ICI therapy		Chemotherapy
	TCM users (n=720)	Non-TCM users (n=1,792)	TCM users (n=51,292)	Non-TCM users (n=78,718)
Mortality risk
Death (No, %)	390 (54.17)	1242 (69.31)	10978 (21.40)	28949 (36.78)
Total follow-up (person-years)	850.795	1483.6027	95092.75	116426.3
Incidence rate per 100 person-years (95% CI)	45.84 (41.40 to 50.62)	83.72 (79.12 to 88.50)	11.54 (11.33 to 11.76)	24.86 (24.58 to 25.15)
Cox proportional hazard model a
Crude HR (95% CI)	0.57 (0.51 to 0.64)		0.48 (0.47 to 0.49)
Adjusted HR (95% CI) b	0.58 (0.51 to 0.66)		0.59 (0.57 to 0.60)

^aThe HR is calculated using the non-TCM users as the reference.

^badjustment for age, sex, cancer type, cancer stage, personal income, calendar year and comorbidities.

ICI, immune checkpoint inhibitor; TCM, Traditional Chinese Medicine.

As shown in Tables 3 and 4, TCM users exhibited lower mortality rates than non-TCM users across all patients who received ICIs or chemotherapy in most different subgroups. In the ICI population (Table 3), TCM users with advanced cancer stages (stages 3 and 4) showed a significantly lower mortality rate than non-TCM users, and similar trends were observed in patients with early-stage cancer. The adjusted HRs were 0.57 and 0.56, respectively. Across different age groups, TCM users consistently showed lower mortality rates, with the largest benefits observed in patients aged 41–65 years and those aged >66 years, where TCM usage was associated with a 41%–43% reduction in mortality risk after adjustments. When analyzed by sex, both men and women who used TCM showed lower mortality rates than those who did not use TCM, with adjusted HRs indicating a significant survival benefit (0.58 and 0.60, respectively). Further analysis of mortality rates associated with several comorbidities showed that, irrespective of whether patients had these conditions, the mortality rate reduced in TCM users. Notably, significant interaction was observed for diabetes mellitus (P for interaction = 0.018) and dyslipidemia (P for interaction = 0.024), indicating that these specific comorbidities may influence the magnitude of the survival benefit associated with TCM use.

Table 3.

Comparison of Mortality Rates between Traditional Chinese Medicine (TCM) Users and Non-TCM users Among Cancer Patients Receiving ICI Treatment across Various Subgroups

Variables	Death		Cohort Study
	TCM	Non-TCM users	Crude HR (95% CI)	Adjusted HR a (95% CI)	P for interaction
	n (%)	n (%)	TCM users vs. Non-TCM users
Subgroups
Cancer Stage
3-4	165 (51.72)	653 (69.84)	0.55 (0.46 to 0.65)	0.57 (0.48 to 0.68)	0.943
0-2	77 (51.68)	226 (67.87)	0.57 (0.44 to 0.73)	0.56 (0.43 to 0.73)
Age group
≤ 40 years old	11 (40.74)	22 (57.89)	0.61 (0.29 to 1.25)	0.63 (0.26 to 1.49)	0.383
41-65 years old	210 (54.69)	605 (68.21)	0.60 (0.52 to 0.71)	0.59 (0.5 to 0.71)
≥ 66 years old	169 (54.69)	615 (70.93)	0.55 (0.46 to 0.65)	0.57 (0.47 to 0.68)
Sex
Men	277 (54.42)	989 (69.7)	0.56 (0.49 to 0.64)	0.58 (0.5 to 0.67)	0.676
Women	113 (53.55)	253 (67.83)	0.61 (0.49 to 0.76)	0.60 (0.47 to 0.78)
Cancer type
Liver	72 (66.06)	176 (80.37)	0.65 (0.49 to 0.86)	0.70 (0.53 to 0.93)	0.100
Lung	82 (43.85)	315 (64.55)	0.50 (0.39 to 0.64)	0.53 (0.41 to 0.68)
Oral	55 (56.12)	261 (72.5)	0.54 (0.4 to 0.72)	0.51 (0.38 to 0.69)
Diabetes mellitus
Yes	83 (50.92)	325 (71.12)	0.50 (0.39 to 0.63)	0.46 (0.35 to 0.6)	0.018
No	307 (55.12)	917 (68.69)	0.60 (0.53 to 0.68)	0.62 (0.54 to 0.72)
Hypertension
Yes	145 (53.51)	497 (68.84)	0.55 (0.46 to 0.67)	0.55 (0.44 to 0.67)	0.227
No	245 (54.57)	745 (69.63)	0.58 (0.51 to 0.68)	0.60 (0.51 to 0.7)
Dyslipidemia
Yes	69 (47.59)	241 (67.89)	0.49 (0.37 to 0.64)	0.43 (0.32 to 0.58)	0.024
No	321 (55.83)	1001 (69.66)	0.59 (0.52 to 0.67)	0.61 (0.53 to 0.7)
Heart failure
Yes	-- b	40 (81.63)	0.99 (0.46 to 2.12)	0.54 (0.16 to 1.81)	0.514
No	382 (53.8)	1202 (68.96)	0.57 (0.51 to 0.64)	0.58 (0.51 to 0.66)
Stroke
Yes	17 (73.91)	51 (76.12)	0.73 (0.42 to 1.27)	0.69 (0.33 to 1.44)	0.884
No	373 (53.52)	1191 (69.04)	0.57 (0.51 to 0.64)	0.58 (0.51 to 0.66)
COPD
Yes	36 (50.7)	143 (71.14)	0.46 (0.32 to 0.66)	0.36 (0.24 to 0.55)	0.109
No	354 (54.55)	1099 (69.08)	0.59 (0.52 to 0.66)	0.61 (0.54 to 0.7)
CKD
Yes	34 (51.52)	98 (63.64)	0.59 (0.4 to 0.87)	0.41 (0.25 to 0.67)	0.495
No	356 (54.43)	1144 (69.84)	0.57 (0.51 to 0.64)	0.59 (0.52 to 0.67)
Depression
Yes	22 (55)	35 (58.33)	0.81 (0.47 to 1.38)	0.74 (0.36 to 1.52)	0.363
No	368 (54.12)	1207 (69.69)	0.57 (0.5 to 0.64)	0.58 (0.51 to 0.65)
Anxiety
Yes	40 (56.34)	80 (73.39)	0.59 (0.4 to 0.87)	0.43 (0.27 to 0.69)	0.752
No	350 (53.93)	1162 (69.04)	0.57 (0.51 to 0.64)	0.59 (0.52 to 0.67)
Anemia
Yes	13 (59.09)	28 (73.68)	0.64 (0.33 to 1.24)	0.51 (0.2 to 1.27)	0.588
No	377 (54.01)	1214 (69.21)	0.57 (0.51 to 0.64)	0.58 (0.51 to 0.66)
Number of visits to TCM
< 2	99 (63.87)	1242 (69.31)	0.81 (0.66 to 1.00)	0.80 (0.64 to 1.00)	<0.001 †
2∼3	84 (58.33)		0.65 (0.52 to 0.81)	0.62 (0.48 to 0.79)
4∼9	126 (57.01)		0.60 (0.50 to 0.72)	0.65 (0.53 to 0.78)
≥ 10	81 (40.5)		0.38 (0.30 to 0.47)	0.39 (0.30 to 0.50)

^aadjustment for age, sex, cancer type, cancer stage, personal income, calendar year and comorbidities.

^bDue to data usage regulations associated with the Taiwan Health Insurance Database, small counts cannot be displayed to ensure compliance with privacy requirements.

^†P for trend.

HR, hazard ratio; CI, confidence intervals.

Table 4.

Comparison of Mortality Rates between Traditional Chinese Medicine (TCM) Users and Non-TCM Users among Cancer Patients Receiving Chemotherapy Treatment across Various Subgroups

Variables	Death		Cohort Study
	TCM	Non-TCM users	Crude HR (95% CI)	Adjusted HR a (95% CI)	P for interaction
	n (%)	n (%)	TCM users vs. Non-TCM users
Subgroups
Cancer Stage
3-4	4912 (34.56)	14383 (51.06)	0.54 (0.53 to 0.56)	0.58 (0.57 to 0.6)	0.008
0-2	2137 (12.1)	5227 (23.06)	0.46 (0.44 to 0.49)	0.57 (0.54 to 0.6)
Age group
≤ 40 years old	407 (7.93)	598 (14.49)	0.50 (0.44 to 0.57)	0.67 (0.58 to 0.77)	<0.001
41-65 years old	5537 (19.64)	12108 (31.63)	0.53 (0.51 to 0.54)	0.64 (0.62 to 0.66)
≥ 66 years old	5034 (28.01)	16243 (44.74)	0.50 (0.49 to 0.52)	0.55 (0.53 to 0.57)
Sex
Men	5933 (27.42)	19423 (40.77)	0.55 (0.54 to 0.57)	0.60 (0.58 to 0.62)	0.119
Women	5045 (17.01)	9526 (30.65)	0.47 (0.46 to 0.49)	0.58 (0.56 to 0.6)
Cancer Type
Liver	891 (37.33)	2389 (49.61)	0.61 (0.57 to 0.66)	0.637 (0.59 to 0.69)	<0.001
Lung	1541 (41.5)	4479 (63.08)	0.49 (0.46 to 0.52)	0.60 (0.56 to 0.63)
Oral	791 (19.49)	2850 (35.36)	0.46 (0.42 to 0.50)	0.52 (0.48 to 0.56)
Diabetes mellitus
Yes	2714 (27.00)	8253 (42.68)	0.51 (0.49 to 0.53)	0.57 (0.55 to 0.6)	0.307
No	8264 (20.04)	20 696 (34.85)	0.48 (0.47 to 0.49)	0.60 (0.58 to 0.62)
Hypertension
Yes	4130 (23.96)	12 577 (40.04)	0.49 (0.47 to 0.50)	0.56 (0.54 to 0.58)	<0.001
No	6848 (20.11)	16 372 (34.61)	0.49 (0.47 to 0.50)	0.62 (0.60 to 0.63)
Dyslipidemia
Yes	2468 (20.46)	6481 (35)	0.49 (0.46 to 0.51)	0.57 (0.55 to 0.60)	0.012
No	8510 (21.69)	22 468 (37.32)	0.48 (0.47 to 0.49)	0.60 (0.58 to 0.61)
Heart failure
Yes	380 (36.68)	1415 (55.23)	0.50 (0.45 to 0.56)	0.56 (0.50 to 0.64)	0.857
No	10 598 (21.09)	27 534 (36.15)	0.48 (0.47 to 0.50)	0.59 (0.58 to 0.61)
Stroke
Yes	408 (30.75)	1429 (49.36)	0.49 (0.44 to 0.55)	0.57 (0.50 to 0.64)	0.026
No	10 570 (21.15)	27 520 (36.3)	0.48 (0.47 to 0.49)	0.59 (0.58 to 0.61)
COPD
Yes	751 (35.41)	2832 (55.74)	0.48 (0.44 to 0.52)	0.56 (0.52 to 0.61)	0.361
No	10 227 (20.8)	26 117 (35.47)	0.49 (0.48 to 0.50)	0.59 (0.58 to 0.61)
CKD
Yes	1004 (27.95)	3350 (47.32)	0.46 (0.43 to 0.49)	0.51 (0.47 to 0.55)	0.001
No	9974 (20.91)	25 599 (35.73)	0.49 (0.48 to 0.50)	0.60 (0.59 to 0.62)
Depression
Yes	623 (24.43)	1264 (41.4)	0.48 (0.43 to 0.52)	0.56 (0.50 to 0.62)	0.692
No	10 355 (21.24)	27 685 (36.59)	0.48 (0.47 to 0.49)	0.59 (0.58 to 0.61)
Anxiety
Yes	1200 (20.97)	2118 (36.89)	0.47 (0.44 to 0.51)	0.57 (0.53 to 0.62)	0.700
No	9778 (21.46)	26 831 (36.77)	0.48 (0.47 to 0.49)	0.60 (0.58 to 0.61)
Anemia
Yes	388 (22.73)	990 (41.42)	0.44 (0.40 to 0.50)	0.54 (0.48 to 0.62)	0.344
No	10 590 (21.36)	27 959 (36.63)	0.48 (0.47 to 0.49)	0.59 (0.58 to 0.61)
Number of visits to TCM
< 2	2590 (25.81)	28 949 (36.78)	0.62 (0.59 to 0.64)	0.70 (0.67 to 0.73)	<0.001 †
2∼5	3930 (25.49)		0.60 (0.58 to 0.62)	0.71 (0.69 to 0.74)
6∼15	2755 (21.73)		0.50 (0.48 to 0.52)	0.63 (0.61 to 0.66)
≥ 16	1703 (12.94)		0.27 (0.25 to 0.28)	0.35 (0.33 to 0.37)

^aadjustment for age, sex, cancer type, cancer stage, personal income, calendar year and comorbidities.

^†P for trend.

HR, hazard ratio; CI, confidence intervals.

Similarly, in the chemotherapy population, the adjusted HRs for TCM users with both advanced and early-stage cancer were 0.58 and 0.57, respectively, suggesting that TCM use is associated with a 42%–43% decrease in mortality risk among chemotherapy-administered patients, irrespective of cancer stage (Table 4). Significant interaction effects were identified for several factors, including cancer stage (P for interaction = 0.008), age (P for interaction < 0.001), cancer type (P for interaction < 0.001), hypertension (P for interaction < 0.001), dyslipidemia (P for interaction = 0.012), stroke (P for interaction = 0.026) and chronic kidney disease (P for interaction = 0.001), demonstrating that while TCM use consistently and significantly reduces mortality risk across all patient groups, the magnitude of this survival benefit is further modified by the patient’s specific clinical profile. Additionally, a dose–response relationship was observed in both cohorts (P for trend < 0.001), with greater reductions in mortality risk associated with higher frequency of TCM use. After adjustment, the protective effect of frequent TCM use remained significant. In the ICI group, individuals with ≥ 10 TCM visits exhibited a 61% lower risk of death (adjusted HR: 0.39, 95% CI: 0.30–0.50). Similarly, in the chemotherapy group, patients with ≥ 16 TCM visits had a 65% lower risk of death (adjusted HR: 0.35, 95% CI: 0.33–0.37).

To gain deeper insights into the mechanisms through which TCM may reduce mortality in patients with cancer, we identified the most frequently used TCM formulas and single herbs, which were further categorized into therapeutic functional groups according to their primary therapeutic actions (Figure 2). The predominant categories included “clear heat and remove toxin” and “tonify Qi and strengthen spleen,” indicating the primary therapeutic strategies adopted in this population. Interestingly, the three most commonly used herbal medicines among patients receiving ICIs or chemotherapy were the same, viz., Hedyotis diffusa Willd (Bai Hua She She Cao), Scutellaria barbata (Ban Zhi Lian), and Astragali Radix (Huang Qi). Other shared herbs included Rhei Radix et Rhizoma (Da Huang), Mongolian dandelion herb (Pu Gong Ying), Scutellariae Radix (Huang Qin), and Magnoliae Cortex (Hou Pu). The Latin name (or English name) and Chinese name for these herbs are shown in Supplementary Table 2.

Figure 2.

Ranking and distribution of traditional Chinese medicine prescriptions in patients receiving ICI or Chemotherapy

Discussion

This study analyzed data from the Taiwan’s NHIRD and showed that patients with cancer receiving ICIs or chemotherapy exhibited significantly lower mortality and incidence rates when using TCM than patients not using TCM. Moreover, TCM users consistently showed lower mortality rates across various subgroups of patients receiving either ICIs or chemotherapy.

To our knowledge, this study represents the first large-scale, real-world population-based investigation on the combined use of TCM and ICIs in cancer care. Although previous population-based studies primarily focused on the impact of TCM in chemotherapy users, studies investigating its effects on novel cancer therapies, such as ICIs, are limited. Most previous studies on this topic have relied on animal or cellular experiments to clarify the mechanisms of action of herbal medicine. ²¹ Although several small-sample clinical trials have been published, and some larger scale clinical studies are ongoing, there remains a lack of large-scale clinical investigations addressing this issue. ¹³ Our findings provide robust evidence from a large, real-world dataset, bridging the existing research gap and demonstrating the potential synergistic benefits of integrating TCM with ICIs or chemotherapy in clinical practice.

In our analysis, TCM use was associated with significant reductions in mortality among ICI-treated patients with various comorbidities. Remarkably, the most pronounced benefit was observed in patients with COPD, where the adjusted HR decreased markedly from 0.61 to 0.36. Similar reductions were detected in patients with diabetes, dyslipidemia, and chronic kidney disease. These comorbidities share common underlying pathophysiological factors, including chronic systemic inflammation, metabolic dysregulation, oxidative stress, and immune dysfunction. ²² Among these conditions, diabetes warrants special attention because of its well-documented effect on cancer prognosis. A systematic review of 97 studies reported a 1.25-fold increased risk of cancer-related mortality in diabetic individuals compared with that in nondiabetic individuals, even after adjusting for confounders. ²³ Despite this increased risk, our findings are consistent with those of previous studies demonstrating that TCM use reduces mortality (HR: 0.783) in diabetic patients with carcinoma in situ. ²⁴ The diagnostic approach of TCM, based on overall symptom patterns, emphasizes holistic regulation of the human body, which may complement the management of multiorgan damage associated with diabetes. ²⁵ Our findings emphasize the therapeutic potential of TCM in improving outcomes in patients with cancer with multifactorial comorbidities receiving ICIs. Future studies should further explore the role of TCM in managing patients with cancer with other chronic disorders, such as chronic kidney disease and dyslipidemia.

We also observed gender-related differences in TCM usage. ICI-treated patients were predominantly male, with TCM use being more common, which may be attributed to the higher prevalence of cancers such as lung, liver, and bladder cancers among men, which are also cancers where ICIs are widely applied. ² Conversely, in the chemotherapy group, a higher proportion of female patients used TCM, a finding consistent with previous NHI surveys, which reported higher annual per capita TCM usage rates among women (1.477 visits) than among men (1.051 visits). ²⁶ These gender-related differences in TCM use may reflect variations in cancer prevalence and healthcare-seeking behaviors, necessitating further investigation.

Regarding cancer stage, we observed that TCM use during chemotherapy was associated with a 42%–43% reduction in mortality risk in patients with both advanced- and early-stage cancer. This result is consistent with previous research reporting a 32% reduction in mortality risk among patients with lung cancer using TCM compared with that among patients not using TCM (HR: 0.62). ²⁷ Similarly, in breast cancer cases, TCM use significantly reduced all-cause mortality, with adjusted HRs of 0.55 for TCM use lasting 30–180 days and 0.46 for TCM use exceeding 180 days. ²⁸ These data emphasize the potential of TCM as an adjunctive therapy to improve survival outcomes across various cancer types and stages. The formal interaction testing conducted in this study provides deeper insights into the heterogeneity of TCM’s therapeutic effects. The significant interactions observed, particularly in the chemotherapy group for factors like age, cancer stage, cancer type, hypertension, dyslipidemia, stroke and chronic kidney disease, suggest that TCM may offer a more pronounced survival advantage for specific patient populations. This highlights the importance of personalized integrative oncology, where TCM interventions are tailored to the patient’s baseline comorbidities and treatment intensity. Furthermore, the highly significant P-trend for TCM visit frequency in both the ICI and chemotherapy cohorts reinforces the hypothesis of a dose-response protective effect. This finding is consistent with a study on patients with gastric cancer receiving surgery and adjuvant chemotherapy, which found that both short-term (<180 days) and long-term (≥180 days) TCM use were independently associated with reduced mortality risk (HR: 0.59 and 0.41, respectively). ¹⁶ These data emphasize the importance of regular TCM use in achieving optimal therapeutic outcomes.

Our study accurately reflects real-world TCM clinical practices, where treatments rarely depend on single herbs or fixed prescriptions. The most frequently used herbal medicines— Hedyotis diffusa Willd (Bai Hua She She Cao), Scutellaria barbata (Ban Zhi Lian), which are primarily used to clear heat and remove toxins, alongside Astragali Radix (Huang Qi), utilized to tonify Qi and strengthen the Spleen—align with the TCM principles of “pathogen-eliminating” (Qu-Xie) and “health-strengthening” (Fu-Zheng), respectively.^28,29 Astragali Radix, a widely recognized health-strengthening herb, contains bioactive components such as astragalus polysaccharides (APS) that regulate cytokines, modulate key signaling pathways (e.g., TLR4, NF-κB, and MAPK), and improve immune cell functions. APS optimize immune balance within the tumor microenvironment, advancing immunotherapy as an adjunct to ICIs.^30,31 Scutellaria barbata and Hedyotis diffusa Willd, known for their pathogen-eliminating properties, exert antitumor effects, including proliferation inhibition, apoptosis induction, angiogenesis suppression, and immune regulation.^32,33 By confirming the immunomodulatory and antitumor benefits of these herbs, our study provides valuable evidence supporting TCM integration into modern cancer treatment. Future research and clinical trials are necessary to further validate these findings, ultimately benefiting patients with cancer globally.

This study has some unique strengths. First, it leverages large-scale claims data from the Taiwan’s NHIRD, ensuring a highly representative sample and minimizing selection bias. Second, the use of real-world clinical practice data provides a more accurate reflection of the actual treatment strategies and patient characteristics compared with controlled clinical trials, providing valuable insights into the practical effectiveness of therapies. Third, our study used a robust design to compare mortality outcomes between patients with cancer receiving ICIs who used TCM and those who did not, yielding promising results. Despite these strengths, this study has several limitations. First, the NHIRD lacks critical variables that may impact outcomes, such as lifestyle factors, nutritional status, and genetic information, potentially influencing the comprehensiveness of the analysis. Second, the Taiwan NHI does not cover CTLA-4 inhibitors, restricting our findings to patients treated with PD-1 and PD-L1 inhibitors, which limits the generalizability of our results to other ICI therapies. Third, the inherent complexity of TCM research poses challenges, particularly due to the principle of “syndrome differentiation and treatment” and the multicomponent, multitarget, and multipathway characteristics of TCM. These factors cause considerable variability in herbal combinations, dosages, and treatment durations, complicating standardization and analysis.

Conclusions

This nationwide study suggests that TCM users have lower mortality than non-TCM users among patients receiving ICIs or chemotherapy. Moreover, the majority of subgroup analyses consistently highlighted the beneficial effects of TCM in improving cancer treatment outcomes. This association persisted even after adjusting for potential confounding factors, emphasizing the robustness of the findings. Our results also emphasize the importance of conducting further prospective studies to confirm these observations and investigate the underlying biological and therapeutic mechanisms. Such research could significantly contribute to the advancement of integrative cancer care, fostering a more comprehensive approach to improving patient outcomes.

ORCID iD

Tsung-Yun Hsieh https://orcid.org/0000-0003-1576-4057

Ethical Considerations

This study was reviewed by the Tri-Service General Hospital Institutional Review Board (TSGHIRB C202305093).

Consent to Participate

Informed consent was waived because the study used de-identified secondary data from a nationwide database.

Data Availability Statement

Data used in this study are handled and stored by the Health and Welfare Data Science Center. Interested researchers can obtain the data through formal application to the Health and Welfare Data Science Center, Department of Statistics, Ministry of Health and Welfare, Taiwan.*