Mortality differences in coronary patients with breast cancer treated with trimetazidine dihydrochloride: potential therapeutic implications Authors: Yuen-Ting Cheng, Chun-Fung Sin, Edmond S. K. Ma, Stephen T. S. Lam, Bernard MY Cheung, Kai-Hang Yiu, Hung-Fat Tse, Yap-Hang Chan Published: 2025 ABSTRACT BackgroundTrimetazidine, an anti-angina drug with metabolic regulatory effects, has been observationally associated with reduced malignancy risk. However, whether it has clinical implications for established breast cancer is unknown.MethodsWe retrospectively studied 2,526 patients with pre-existing or de novo breast cancer and ischemic heart disease (IHD) who received trimetazidine therapy or oral nitrate, identified from the Hong Kong Clinical Data Analysis and Reporting System (CDARS) between January 1, 1999, and December 31, 2020. Subjects were classified as trimetazidine users (n=169) versus non-users (n=2,357). Multivariable Cox regression was used to estimate the hazard ratio (HR) and 95% confidence interval (CI) of all-cause mortality in the pre-existing breast cancer group.ResultsAfter a mean follow-up duration of 4,235± 2,272 days, patients with pre-existing breast cancer who received trimetazidine experienced had a longer overall survival from all-cause mortality than the control group (Mean survival, days [95%CI]; Control, 4,543 [4,163-4,923]; Trimetazidine, 5,405 [2,857-7,953]; Breslow [Generalised Wilcoxon], 5.07, p=.02). Cox regression showed that use of trimetazidine was independently associated with lower all-cause mortality (adjusted HR, 0.48; 95% CI, 0.48-0.98; P = .04). Sensitivity analyses analyzing subjects with pre-existing or de novo breast cancer yielded similar findings (adjusted HR, 0.75; 95% CI, 0.60-0.94; P = .01). The survival benefits were mainly conferred to non-cardiovascular mortality (Breslow [Generalised Wilcoxon], 10.6, p=.001).ConclusionsTrimetazidine use is associated with a lower risk of all-cause mortality among breast cancer patients with IHD. While causality cannot be established at present, further randomized controlled trials are warranted to confirm our findings and further elucidate the underlying mechanisms.Supplementary InformationThe online version contains supplementary material available at 10.1186/s40959-025-00433-3. Introduction Breast cancer is the most common cancer [1] and the leading cause of mortality in females [2, 3]. Breast cancer’s global incidence is estimated to reach around 3.2 million per year by 2050 [4]. While the importance of public health prevention and risk factors control cannot be overstated, breast cancer patients remain challenged with limited treatment options, especially those with advanced diseases. Conventional chemotherapeutic agents and newer therapeutic options, including immunotherapies, have their inherent or cost-related limitations. There is a need to search for novel therapies that act on unconventional mechanisms, and that are also financially affordable to patients. In this regard, metabolic reprogramming represents a well-described series of cellular metabolic processes that are involved in cancer cell proliferation [5], metastases [6], and the development of drug resistance [7]. In breast cancer, metabolic reprogramming is regulated intrinsically by metabolic regulators, such as p53 [8], MYC [9]. Cancer cells also shift to alternative pathways to sustain growth under hypoxia and oxidative stress, such as elevated fatty acid oxidation (FAO) in triple-negative breast cancer (TNBC). Therefore, targeting metabolic pathways may provide a potentially novel avenue for therapeutic interventions. Trimetazidine, an anti-anginal medication for ischemic heart disease (IHD), works by shifting fatty acid metabolism to glucose metabolism [10, 11]. It exerts anti-ischemic [12], anti-inflammatory [13], and anti-apoptotic effects [14]. Recently, trimetazidine has been shown in animal models to exert potential cancer therapeutic effects, alone [15] or in combination with existing chemotherapeutic agent [16]. We recently showed that trimetazidine use is associated with reduced risk of mortality in patients with lung cancer [17]. Nevertheless, relevant clinical data from human studies on specifically breast cancer were lacking. In this study, we sought to investigate whether trimetazidine use may have a clinical impact on patients with breast cancer, in terms of overall survival. We hypothesize based on a modulation hypothesis of metabolic reprogramming, that the use of trimetazidine in IHD patients with breast cancer may result in improved clinical survival reflected in all-cause mortality reduction. Methods Data source This study is a territory-wide, multi-centre, retrospective cohort study that studied clinical data from the Clinical Data Analysis and Reporting System (CDARS) coordinated under the Hong Kong Hospital Authority. CDARS included clinical data from all 42 public hospitals in Hong Kong, that covered catchment population of 7.5 million and provided 90% of population coverage in public clinical services utilization [18]. Research findings generated from CDARS have informed scientific knowledge from prior studies likewise from our group [19] and others. Study design Clinical data on a retrospective cohort of 279,894 patients with IHD who were administered anti-anginal therapies (trimetazidine users, versus non-users as defined by long-acting nitrate use as control) was collected. The study period was defined as from January 1, 1999, to December 31, 2020. We excluded subjects younger than 18 years, those with incomplete sociodemographic or medical history such as sex, diagnosis date of IHD and malignancy (n=1,895), duplications (n=13,751), as well as those with drug exposure of interest less than 30 days (n=55,198). Amongst 20,9050 preliminarily eligible patients with IHD, 2,526 with a diagnosis of breast cancer were identified. Diagnosis of breast cancer was defined according to the International Classification of Diseases (Supplementary Table S1). Pre-existing breast cancer was defined as a breast cancer diagnosis before drug exposure commencement (n=1,012). De novo breast cancer was defined as a breast cancer diagnosis that occurred after drug exposure (n=1,514). Cohort entry was defined as the date of breast cancer diagnosis for the pre-existing group, and the date of first prescription of trimetazidine or long-acting nitrate for the de novo group. For subjects receiving both treatments, the earlier prescription date was used. All participants were followed up from the cohort entry to death, or the right censoring date (December 31, 2020), whichever occurred first. The primary analysis focused on the pre-existing group, observed from breast cancer diagnosis until death or right-censoring for non-events. We also investigated all subjects, including both the pre-existing group and the de novo group, from their respective cohort entry until death or right-censoring for non-events. Definition of trimetazidine exposure The use of trimetazidine was defined as its continuous prescription throughout the study period, with a minimum duration of 30 days or longer. Trimetazidine non-users (controls) were defined as individuals with ischemic heart disease (IHD), who received long-acting nitrate therapy for a minimum duration of 30 days or longer but were not prescribed trimetazidine. Long-acting nitrate therapy was chosen as the control because it is a commonly used first-line anti-anginal agent for the treatment of angina in patients with IHD and has a similar clinical indication to trimetazidine. The clinical outcomes were compared between trimetazidine users (n=169) versus non-users (n=2,357) (Flowchart, Supplementary Figure S1). Intention-to-treat principle was adopted for the current analyses. Definition of primary endpoint The primary endpoint was defined as all-cause mortality of breast cancer patients with trimetazidine use. Secondary endpoints included the relation of trimetazidine use with non-cardiovascular and cardiovascular mortality. Ethics approval and consent All research was performed in accordance with relevant guidelines and regulations. Our study protocol was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (IRB Reference Number: UW-23095). As this is a retrospective observational cohort study, obtaining informed consent was not possible and the need for consent by participants has been waived in the Ethics approval above-stated. Statistical analysis We reported baseline characteristics as frequencies (percentages) for categorical variables and mean (standard deviation [SD]) for continuous variables. Cox proportional hazard regression models were employed to assess hazard ratios (HRs) and 95% confidence intervals (CIs) for all-cause mortality in patients with pre-existing breast cancer. Potential confounders included age, sex, socioeconomic status, cohort recruitment period, chronic smoking exposure (proxied by diagnoses of chronic obstructive pulmonary disease), chronic liver and kidney diseases, and history of cardiovascular disease diagnoses other than IHD (hypertension, Type 2 diabetes mellitus, myocardial infarction, ischemic stroke). Congestive heart failure was used as a proxy for the functional status of patients. Confounding factors in terms of drug use included aspirin and antiplatelet therapy, lipid-lowering therapy, as well as hormone replacement therapy. Further disease-specific confounding factors included advanced stages of disease defined as the presence of metastasis, anatomical location of breast cancer, as well as current guideline-directed treatment modalities: chemotherapeutic agents, tamoxifen, aromatase inhibitor, immune checkpoint inhibitor, targeted therapy, CDK4/6 inhibitor, radiotherapy, and surgical operation (Supplementary Tables S2-4). The analysis includes a 1) crude model, 2) potential confounders with a P-value ≤ 0.10 in the first model, and 3) all potential confounders. Sensitivity analyses were performed in pre-existing or de novo breast cancer patients. We also performed pre-specified analyses on cardiovascular and non-cardiovascular mortality. We classified subjects with ICD-9 death codes of 390 to 459 as cardiovascular mortality, while those with other codes were categorized as non-cardiovascular mortality. All data were analyzed by using the IBM SPSS/PC statistical package Version 29.0. A P value of <.05 was considered statistically significant. Results Baseline clinical characteristics Clinical characteristics of breast cancer patients (n=2,526) are presented in Table 1. Trimetazidine users in our cohort had largely similar clinical characteristics compared to non-users, except that there were more male patients in the trimetazidine group (p<.001). More trimetazidine users had general or cardiovascular comorbidities, especially myocardial infarction (p=.04), and received aspirin/antiplatelet therapy (p=.01) and lipid-lowering therapy (p<.001).Table 1Baseline clinical characteristics of patients with pre-existing or de novo breast cancer (n=2526)aControln=2357Trimetazidinen=169P-valueDemographics Age, years, mean (SD)69.7 (11.2)69.2 (11.0).27 Females2327 (98.7)124 (73.4)<.001 Socioeconomic indicatorb150 (6.4)10 (5.9).82Recruitment period.03 1999-2000867 (36.8)79 (46.7)N/A 2001-20101144 (48.5)66 (39.1)N/A 2011-2020346 (14.7)24 (14.2)N/A Chronic smokingc39 (1.7)1 (0.6).29General/Cardiovascular Comorbidities Type 2 Diabetes Mellitus378 (16.1)35 (20.7).11 Chronic liver diseased14 (0.6)1 (0.6)1.00 Chronic kidney disease53 (2.2)5 (3.0).55 Hypertension606 (25.7)44 (26.0).93 Acute myocardial infarction53 (2.2)8 (4.7).04 Heart failure184 (7.8)14 (8.3).82 Ischemic stroke70 (3.0)5 (3.0).99Breast Cancer CharacteristicsBreast cancer location.36 Axillary tail7 (0.3)0 (0)N/A Lower outer66 (2.8)7 (4.1)N/A Upper outer316 (13.4)14 (8.3)N/A Lower inner52 (2.2)1 (0.6)N/A Upper inner132 (5.6)9 (5.3)N/A Central121 (5.1)6 (3.6)N/A Nipple and areola45 (1.9)2 (1.2)N/A Unspecified1618 (68.6)130 (76.9)N/A Hormone replacement therapye280 (11.9)26 (15.4).18 Surgical Resection of breast cancerf1364 (57.8)98 (58.0).98 Mastectomy1153 (48.9)87 (51.5).52 Advanced breast cancer – presence of metastasisg182 (7.7)15 (8.9).59 Radiotherapyh440 (18.7)35 (20.7).51 Disease-Specific Medications1855 (78.6)138 (81.7).36 Tamoxifeni1271 (53.9)92 (54.4).90 Aromatase inhibitorj903 (38.3)69 (40.8).52 Targeted Therapyk53 (2.2)2 (1.2).36 CDK4/6 Inhibitorl7 (0.3)0 (0).48 Chemotherapym542 (23.0)47 (27.8).15 Immune Checkpoint Inhibitorn8 (0.3)0 (0).45Concurrent Anti-inflammatory agents Aspirin/antiplatelet therapyo2072 (87.9)160 (94.7).01 Lipid-lowering therapyp1451 (61.6)149 (88.2)<.001Values were expressed as number of cases (percentage)aBaseline in patients with pre-existing cancer was defined as the first date of breast cancer diagnosis, and that with de novo breast cancer was defined as the earliest long-acting nitrate or trimetazidine prescription datebSocioeconomic indicator was defined as having received financial support from the comprehensive social security allowance (CSSA) schemecProxied by chronic obstructive pulmonary disease (COPD)dChronic liver disease served as the proximity of the drinking historyeIncluded the use of conjugated estrogen plus medroxyprogesterone acetate (MPA) (Brand names: Prempro, Premphase), Estradiol plus norethisterone acetate (Brand names: Activella, Angeliq), Estradiol plus drospirenone (Brand name: FemHRT), and further search on the hormone replacement therapy procedure (ICD-9 code: 99.24) was usedfICD-9 procedure code of 85.2–85.4.2.4 was employed for all the operations of breast (ICD-9 code of mastectomy: 85.4)gMetastasis to other sits were recorded when the date of metastasis were after the first diagnosis of breast cancerhICD-9 procedure code of 92.2 was employed for all the radiotherapyiIncluded Tamoxifen, Nolvadex and SoltamoxjIncluded Anastrozole (Brand name: Arimidex), Letrozole (Brand name: Femara), and Exemestane (Brand name: Aromasin)kIncluded Trastuzumab (Brand names: Herceptin, Ogivri, Herzuma), Pertuzumab (Brand name: Perjeta), and Ado-trastuzumab emtansine (Brand name: Kadcyla)lIncluded Palbociclib (Brand name: Ibrance), Ribociclib (Brand name: Kisqali), and Abemaciclib (Brand name: Verzenio)mIncluded Anthracycline (Doxorubicin/Adriamycin, Epirubicin/Ellence), Taxanes (Paclitaxel/Taxol, Docetaxel/Taxotere), Cyclophosphamide (Cytoxan), Fluorouracil (5-FU), Carboplatin, and Gemcitabine (Gemzar)nIncluded pembrolizumab (Brand name: Keytruda)oIncluded aspirin, clopidogrel, ticagrelor, and prasugrelpIncluded simvastatin, rosuvastatin, pravastatin, fluvastatin, atorvastatin, ezetimibe, fenofibrate, alirocumab, and evolocumab Table 1Baseline clinical characteristics of patients with pre-existing or de novo breast cancer (n=2526)aControln=2357Trimetazidinen=169P-valueDemographics Age, years, mean (SD)69.7 (11.2)69.2 (11.0).27 Females2327 (98.7)124 (73.4)<.001 Socioeconomic indicatorb150 (6.4)10 (5.9).82Recruitment period.03 1999-2000867 (36.8)79 (46.7)N/A 2001-20101144 (48.5)66 (39.1)N/A 2011-2020346 (14.7)24 (14.2)N/A Chronic smokingc39 (1.7)1 (0.6).29General/Cardiovascular Comorbidities Type 2 Diabetes Mellitus378 (16.1)35 (20.7).11 Chronic liver diseased14 (0.6)1 (0.6)1.00 Chronic kidney disease53 (2.2)5 (3.0).55 Hypertension606 (25.7)44 (26.0).93 Acute myocardial infarction53 (2.2)8 (4.7).04 Heart failure184 (7.8)14 (8.3).82 Ischemic stroke70 (3.0)5 (3.0).99Breast Cancer CharacteristicsBreast cancer location.36 Axillary tail7 (0.3)0 (0)N/A Lower outer66 (2.8)7 (4.1)N/A Upper outer316 (13.4)14 (8.3)N/A Lower inner52 (2.2)1 (0.6)N/A Upper inner132 (5.6)9 (5.3)N/A Central121 (5.1)6 (3.6)N/A Nipple and areola45 (1.9)2 (1.2)N/A Unspecified1618 (68.6)130 (76.9)N/A Hormone replacement therapye280 (11.9)26 (15.4).18 Surgical Resection of breast cancerf1364 (57.8)98 (58.0).98 Mastectomy1153 (48.9)87 (51.5).52 Advanced breast cancer – presence of metastasisg182 (7.7)15 (8.9).59 Radiotherapyh440 (18.7)35 (20.7).51 Disease-Specific Medications1855 (78.6)138 (81.7).36 Tamoxifeni1271 (53.9)92 (54.4).90 Aromatase inhibitorj903 (38.3)69 (40.8).52 Targeted Therapyk53 (2.2)2 (1.2).36 CDK4/6 Inhibitorl7 (0.3)0 (0).48 Chemotherapym542 (23.0)47 (27.8).15 Immune Checkpoint Inhibitorn8 (0.3)0 (0).45Concurrent Anti-inflammatory agents Aspirin/antiplatelet therapyo2072 (87.9)160 (94.7).01 Lipid-lowering therapyp1451 (61.6)149 (88.2)<.001Values were expressed as number of cases (percentage)aBaseline in patients with pre-existing cancer was defined as the first date of breast cancer diagnosis, and that with de novo breast cancer was defined as the earliest long-acting nitrate or trimetazidine prescription datebSocioeconomic indicator was defined as having received financial support from the comprehensive social security allowance (CSSA) schemecProxied by chronic obstructive pulmonary disease (COPD)dChronic liver disease served as the proximity of the drinking historyeIncluded the use of conjugated estrogen plus medroxyprogesterone acetate (MPA) (Brand names: Prempro, Premphase), Estradiol plus norethisterone acetate (Brand names: Activella, Angeliq), Estradiol plus drospirenone (Brand name: FemHRT), and further search on the hormone replacement therapy procedure (ICD-9 code: 99.24) was usedfICD-9 procedure code of 85.2–85.4.2.4 was employed for all the operations of breast (ICD-9 code of mastectomy: 85.4)gMetastasis to other sits were recorded when the date of metastasis were after the first diagnosis of breast cancerhICD-9 procedure code of 92.2 was employed for all the radiotherapyiIncluded Tamoxifen, Nolvadex and SoltamoxjIncluded Anastrozole (Brand name: Arimidex), Letrozole (Brand name: Femara), and Exemestane (Brand name: Aromasin)kIncluded Trastuzumab (Brand names: Herceptin, Ogivri, Herzuma), Pertuzumab (Brand name: Perjeta), and Ado-trastuzumab emtansine (Brand name: Kadcyla)lIncluded Palbociclib (Brand name: Ibrance), Ribociclib (Brand name: Kisqali), and Abemaciclib (Brand name: Verzenio)mIncluded Anthracycline (Doxorubicin/Adriamycin, Epirubicin/Ellence), Taxanes (Paclitaxel/Taxol, Docetaxel/Taxotere), Cyclophosphamide (Cytoxan), Fluorouracil (5-FU), Carboplatin, and Gemcitabine (Gemzar)nIncluded pembrolizumab (Brand name: Keytruda)oIncluded aspirin, clopidogrel, ticagrelor, and prasugrelpIncluded simvastatin, rosuvastatin, pravastatin, fluvastatin, atorvastatin, ezetimibe, fenofibrate, alirocumab, and evolocumab Baseline clinical characteristics of patients with pre-existing or de novo breast cancer (n=2526)a Values were expressed as number of cases (percentage) aBaseline in patients with pre-existing cancer was defined as the first date of breast cancer diagnosis, and that with de novo breast cancer was defined as the earliest long-acting nitrate or trimetazidine prescription date bSocioeconomic indicator was defined as having received financial support from the comprehensive social security allowance (CSSA) scheme cProxied by chronic obstructive pulmonary disease (COPD) dChronic liver disease served as the proximity of the drinking history eIncluded the use of conjugated estrogen plus medroxyprogesterone acetate (MPA) (Brand names: Prempro, Premphase), Estradiol plus norethisterone acetate (Brand names: Activella, Angeliq), Estradiol plus drospirenone (Brand name: FemHRT), and further search on the hormone replacement therapy procedure (ICD-9 code: 99.24) was used fICD-9 procedure code of 85.2–85.4.2.4 was employed for all the operations of breast (ICD-9 code of mastectomy: 85.4) gMetastasis to other sits were recorded when the date of metastasis were after the first diagnosis of breast cancer hICD-9 procedure code of 92.2 was employed for all the radiotherapy iIncluded Tamoxifen, Nolvadex and Soltamox jIncluded Anastrozole (Brand name: Arimidex), Letrozole (Brand name: Femara), and Exemestane (Brand name: Aromasin) kIncluded Trastuzumab (Brand names: Herceptin, Ogivri, Herzuma), Pertuzumab (Brand name: Perjeta), and Ado-trastuzumab emtansine (Brand name: Kadcyla) lIncluded Palbociclib (Brand name: Ibrance), Ribociclib (Brand name: Kisqali), and Abemaciclib (Brand name: Verzenio) mIncluded Anthracycline (Doxorubicin/Adriamycin, Epirubicin/Ellence), Taxanes (Paclitaxel/Taxol, Docetaxel/Taxotere), Cyclophosphamide (Cytoxan), Fluorouracil (5-FU), Carboplatin, and Gemcitabine (Gemzar) nIncluded pembrolizumab (Brand name: Keytruda) oIncluded aspirin, clopidogrel, ticagrelor, and prasugrel pIncluded simvastatin, rosuvastatin, pravastatin, fluvastatin, atorvastatin, ezetimibe, fenofibrate, alirocumab, and evolocumab Association of trimetazidine use with lower all-cause mortality in patients with breast cancer Over the mean follow-up duration of 4,235± 2,272 days, death occurred in 1,467 patients (58.1%) in the cohort. In patients with pre-existing breast cancer, there was a total of 609 (60.2%) death events, including 35 (50.7%) trimetazidine users and 574 controls (60.9%) (Supplementary Table S5). The follow-up time from breast cancer diagnosis to first drug prescription was 1960± 1753 days. Trimetazidine users had a longer event-free survival from all-cause mortality compared to control (Mean survival, days [95%CI]; Control, 4,543 [4,163-4,923]; Trimetazidine, 5,405 [2,857-7,953]; Breslow [Generalised Wilcoxon], 5.07, p=.02, Fig. 1). Using Cox proportional hazards regression, trimetazidine use was associated with a lower risk of all-cause mortality than control (crude HR, 0.65; 95% CI, 0.47–0.92.47.92; P =.02; Table 2). Such finding remained persistent after adjustment for potential confounders (adjusted HR, 0.48; 95% CI, 0.48–0.98.48.98; P =.04; Table 2). Sensitivity analysis of subjects with pre-existing or de novo breast cancers revealed a similar risk reduction of all-cause mortality with trimetazidine use (adjusted HR, 0.75; 95% CI, 0.60–0.94.60.94; P =.01; Table 3). Trimetazidine users also had a longer event-free survival from all-cause mortality compared to control (Mean survival, days [95%CI]; Control, 4,950 [4,839-5,061]; Trimetazidine, 5,531 [5,159-5,902]; Breslow [Generalised Wilcoxon], 8.87, p=.003, Fig. 2).Fig. 1Kaplan-meier curve analyses for all-cause mortality in patients with pre-existing breast cancers. Trimetazidine group had a longer event-free survival from all-cause mortality compared to control (Mean survival, days [95%CI]; Control, 4,543 [4,163-4,923]; Trimetazidine, 5,405 [2,857-7,953]; Breslow [Generalised Wilcoxon], 5.07, p=.02)Table 2Predictors for all-cause mortality in patients with pre-existing breast cancers (n=1012)aCrude modelMultivariable model 1Multivariable model 2HR [95% CI]P-valueHR [95% CI]P-valueHR [95% CI]P-valueDemographics Age1.04 [1.03 to 1.04]<.001*1.06 [1.05 to 1.07]<.001*1.03 [1.02 to 1.04]<.001* Male0.93 [0.59 to 1.46].74N/AN/A1.45 [0.83 to 2.54].19 Socioeconomic indicatorb0.64 [0.43 to 0.94].02*0.53 [0.35 to 0.79].002*0.56 [0.37 to 0.83].005* Chronic smokingc3.76 [2.37 to 5.96]<.001*1.68 [0.99 to 2.84].062.01 [1.18 to 3.43].01*General/Cardiovascular Comorbidities Type 2 Diabetes Mellitus1.94 [1.58 to 2.38]<.001*1.53 [1.19 to 1.96]<.001*1.51 [1.18 to 1.94].001* Chronic liver diseased1.34 [0.43 to 4.17].62N/AN/A1.72 [0.54 to 5.49].36 Chronic kidney disease3.38 [2.29 to 4.99]<.001*3.38 [2.18 to 5.21]<.001*3.03 [1.97 to 4.67]<.001* Acute myocardial infarction5.64 [2.89 to 11.0]<.001*5.28 [2.68 to 10.4]<.001*6.26 [3.17 to 12.4]<.001* Heart failure4.07 [2.89 to 5.73]<.001*1.53 [1.03 to 2.25].03*1.90 [1.29 to 2.81].001* Hypertension1.83 [1.53 to 2.19]<.001*1.21 [0.97 to 1.51].101.35 [1.08 to 1.69].01* Ischemic stroke1.90 [1.17 to 3.09].009*1.91 [1.16 to 3.15].011.97 [1.19 to 3.26].008*Breast Cancer Characteristics Hormone replacement therapye0.90 [0.70 to 1.15].39N/AN/A0.98 [0.75 to 1.28].87 Surgical Resection of breast cancerf0.73 [0.62 to 0.86]<.001*0.82 [0.69 to 0.98].030.81 [0.68 to 0.96].02* Presence of metastasisg2.45 [1.96 to 3.06]<.001*2.43 [1.91 to 3.10]<.001*2.50 [1.93 to 3.22]<.001* Radiotherapyh0.87 [0.71 to 1.07].18N/AN/A1.09 [0.86 to 1.38].47 Tamoxifeni0.99 [0.84 to 1.16].89N/AN/A0.88 [0.73 to 1.04].14 Aromatase inhibitorj1.15 [0.96 to 1.39].13N/AN/A1.18 [0.96 to 1.46].11 Targeted Therapyk0.70 [0.29 to 1.70].44N/AN/A1.13 [0.45 to 2.88].79 CDK4/6 Inhibitorl0.05 [0 to 18.9].32N/AN/AN/AN/A Chemotherapym0.47 [0.37 to 0.59]<.001*1.07 [0.82 to 1.41].620.73 [0.56 to 0.97].03* Immune Checkpoint Inhibitorn0.67 [0.17 to 2.70].58N/AN/A0.90 [0.22 to 3.71].88Concurrent anti-inflammatory agents Aspirin/antiplatelet therapyo0.51 [0.41 to 0.62]<.001*0.87 [0.69 to 1.10]0.240.87 [0.69 to 1.09].22 Lipid-lowering therapyp0.31 [0.26 to 0.36]<.001*0.48 [0.39 to 0.57]<.001*0.40 [0.33 to 0.49]<.001* Trimetazidine0.65 [0.47 to 0.92].02*0.48 [0.48 to 0.98].04*0.62 [0.40 to 0.95].03**p≤0.05aHazard ratio (HR) prediction estimates and 95% confidence interval explained by variable of interest as shown by univariable and multivariable Cox proportional hazards regression; Crude Model: unadjusted; Multivariable Model 1: adjusted for potential confounders with P- value ≤0.10 in Crude Model; Multivariable Model 2: adjusted for all potential confounders as defined a prioribSocioeconomic indicator was defined as having received financial support from the comprehensive social security allowance (CSSA) schemecProxied by chronic obstructive pulmonary disease (COPD)dChronic liver disease served as the proximity of the drinking historyeIncluded the use of conjugated estrogen plus medroxyprogesterone acetate (MPA) (Brand names: Prempro, Premphase), Estradiol plus norethisterone acetate (Brand names: Activella, Angeliq), Estradiol plus drospirenone (Brand name: FemHRT), and further search on the hormone replacement therapy procedure (ICD-9 code: 99.24) was usedfICD-9 procedure code of 85.2–85.4.2.4 was employed for all the operations of breast (ICD-9 code of mastectomy: 85.4)gMetastasis to other sits were recorded when the date of metastasis were after the first diagnosis of breast cancerhICD-9 procedure code of 92.2 was employed for all the radiotherapyiIncluded Tamoxifen, Nolvadex and SoltamoxjIncluded Anastrozole (Brand name: Arimidex), Letrozole (Brand name: Femara), and Exemestane (Brand name: Aromasin)kIncluded Trastuzumab (Brand names: Herceptin, Ogivri, Herzuma), Pertuzumab (Brand name: Perjeta), and Ado-trastuzumab emtansine (Brand name: Kadcyla)lIncluded Palbociclib (Brand name: Ibrance), Ribociclib (Brand name: Kisqali), and Abemaciclib (Brand name: Verzenio)mIncluded Anthracycline (Doxorubicin/Adriamycin, Epirubicin/Ellence), Taxanes (Paclitaxel/Taxol, Docetaxel/Taxotere), Cyclophosphamide (Cytoxan), Fluorouracil (5-FU), Carboplatin, and Gemcitabine (Gemzar)nIncluded pembrolizumab (Brand name: Keytruda)oIncluded aspirin, clopidogrel, ticagrelor, and prasugrelpIncluded simvastatin, rosuvastatin, pravastatin, fluvastatin, atorvastatin, ezetimibe, fenofibrate, alirocumab, and evolocumabTable 3Predictors for all-cause mortality in patients with pre-existing or de novo breast cancers (n=2526)aCrude modelMultivariable model 1Multivariable model 2HR [95% CI]P-valueHR [95% CI]P-valueHR [95% CI]P-valueDemographics Age1.08 [1.08 to 1.09]<.001*1.05 [1.04 to 1.05]<.001*1.05 [1.04 to 1.05]<.001* Male0.92 [0.68 to 1.24].59N/AN/A0.93 [0.66 to 1.32].69 Socioeconomic indicatorb0.73 [0.58 to 0.91].005*0.63 [0.50 to 0.79]<.001*0.63 [0.50 to 0.80]<.001* Chronic smokingc2.48 [1.76 to 3.49]<.001*1.66 [1.11 to 2.49].01*1.69 [1.13 to 2.52].01*General/Cardiovascular Comorbidities Type 2 Diabetes Mellitus1.99 [1.75 to 2.27]<.001*1.55 [1.32 to 1.81]<.001*1.54 [1.31 to 1.80]<.001* Chronic liver diseased1.43 [0.68 to 3.00].35N/AN/A1.60 [0.71 to 3.60].26 Chronic kidney disease3.19 [2.41 to 4.23]<.001*3.38 [2.50 to 4.58]<.001*3.40 [2.51 to 4.61]<.001* Acute myocardial infarction1.90 [1.40 to 2.58]<.001*2.49 [1.78 to 3.47]<.001*2.46 [1.76 to 3.44]<.001* Heart failure3.00 [2.50 to 3.48]<.001*1.54 [1.28 to 1.86]<.001*1.54 [1.28 to 1.86]<.001* Hypertension1.68 [1.50 to 1.89]<.001*1.22 [1.06 to 1.41].007*1.22 [1.05 to 1.40].007* Ischemic stroke2.07 [1.58 to 2.69]<.001*1.97 [1.48 to 2.63]<.001*1.98 [1.48 to 2.64]<.001*Breast Cancer Characteristics Hormone replacement therapye0.87 [0.74 to 1.02].090.99 [0.83 to 1.18].890.98 [0.82 to 1.16].79 Surgical Resection of breast cancerf0.62 [0.56 to 0.68]<.001*0.76 [0.68 to 0.85]<.001*0.76 [0.67 to 0.85]<.001* Presence of metastasisg2.73 [2.37 to 3.14]<.001*2.84 [2.43 to 3.32]<.001*2.82 [2.41 to 3.30]<.001* Radiotherapyh0.93 [0.82 to 1.07].31N/AN/A1.07 [0.92 to 1.25].38 Tamoxifeni1.29 [1.16 to 1.43]<.001*1.00 [0.89 to 1.12].981.00 [0.89 to 1.12].99 Aromatase inhibitorj0.79 [0.70 to 0.88]<.001*0.88 [0.78 to 0.99].04*0.87 [0.77 to 0.99].03* Targeted Therapyk0.77 [0.48 to 1.23]<.001*1.45 [0.87 to 2.40].161.44 [0.87 to 2.40].16 CDK4/6 Inhibitorl0.43 [0.11 to 1.72].22N/AN/A0.64 [0.16 to 2.61].54 Chemotherapym0.43 [0.37 to 0.50]<.001*0.72 [0.60 to 0.85]<.001*0.72 [0.60 to 0.85]<.001* Immune Checkpoint Inhibitorn0.70 [0.26 to 1.87].48N/AN/A1.00 [0.37 to 2.70]1.00Concurrent anti-inflammatory agents Aspirin/antiplatelet therapyo0.33 [0.30 to 0.37]<.001*0.89 [0.76 to 1.05]0.170.90 [0.76 to 1.05].18 Lipid-lowering therapyp0.54 [0.47 to 0.62]<.001*0.48 [0.43 to 0.55]<.001*0.48 [0.43 to 0.55]<.001* Trimetazidine0.77 [0.62 to 0.96].02*0.75 [0.60 to 0.94].01*0.76 [0.59 to 0.97].03**p≤0.05aHazard ratio (HR) prediction estimates and 95% confidence interval explained by variable of interest as shown by univariable and multivariable Cox proportional hazards regression; Crude Model: unadjusted; Multivariable Model 1: adjusted for potential confounders with P- value ≤0.10 in Crude Model; Multivariable Model 2: adjusted for all potential confounders as defined a prioribSocioeconomic indicator was defined as having received financial support from the comprehensive social security allowance (CSSA) schemecProxied by chronic obstructive pulmonary disease (COPD)dChronic liver disease served as the proximity of the drinking historyeIncluded the use of conjugated estrogen plus medroxyprogesterone acetate (MPA) (Brand names: Prempro, Premphase), Estradiol plus norethisterone acetate (Brand names: Activella, Angeliq), Estradiol plus drospirenone (Brand name: FemHRT), and further search on the hormone replacement therapy procedure (ICD-9 code: 99.24) was usedfICD-9 procedure code of 85.2–85.4.2.4 was employed for all the operations of breast (ICD-9 code of mastectomy: 85.4)gMetastasis to other sits were recorded when the date of metastasis were after the first diagnosis of breast cancerhICD-9 procedure code of 92.2 was employed for all the radiotherapyiIncluded Tamoxifen, Nolvadex and SoltamoxjIncluded Anastrozole (Brand name: Arimidex), Letrozole (Brand name: Femara), and Exemestane (Brand name: Aromasin)kIncluded Trastuzumab (Brand names: Herceptin, Ogivri, Herzuma), Pertuzumab (Brand name: Perjeta), and Ado-trastuzumab emtansine (Brand name: Kadcyla)lIncluded Palbociclib (Brand name: Ibrance), Ribociclib (Brand name: Kisqali), and Abemaciclib (Brand name: Verzenio)mIncluded Anthracycline (Doxorubicin/Adriamycin, Epirubicin/Ellence), Taxanes (Paclitaxel/Taxol, Docetaxel/Taxotere), Cyclophosphamide (Cytoxan), Fluorouracil (5-FU), Carboplatin, and Gemcitabine (Gemzar)nIncluded pembrolizumab (Brand name: Keytruda)oIncluded aspirin, clopidogrel, ticagrelor, and prasugrelpIncluded simvastatin, rosuvastatin, pravastatin, fluvastatin, atorvastatin, ezetimibe, fenofibrate, alirocumab, and evolocumabFig. 2Kaplan-meier curve analyses for all-cause mortality in patients with pre-existing or de novo breast cancers. Trimetazidine group had a longer event-free survival from all-cause mortality compared to control (Mean survival, days [95%CI]; Control, 4,950 [4,839-5,061]; Trimetazidine, 5,531 [5,159-5,902]; Breslow [Generalised Wilcoxon], 8.87, p=.003) Kaplan-meier curve analyses for all-cause mortality in patients with pre-existing breast cancers. Trimetazidine group had a longer event-free survival from all-cause mortality compared to control (Mean survival, days [95%CI]; Control, 4,543 [4,163-4,923]; Trimetazidine, 5,405 [2,857-7,953]; Breslow [Generalised Wilcoxon], 5.07, p=.02) Table 2Predictors for all-cause mortality in patients with pre-existing breast cancers (n=1012)aCrude modelMultivariable model 1Multivariable model 2HR [95% CI]P-valueHR [95% CI]P-valueHR [95% CI]P-valueDemographics Age1.04 [1.03 to 1.04]<.001*1.06 [1.05 to 1.07]<.001*1.03 [1.02 to 1.04]<.001* Male0.93 [0.59 to 1.46].74N/AN/A1.45 [0.83 to 2.54].19 Socioeconomic indicatorb0.64 [0.43 to 0.94].02*0.53 [0.35 to 0.79].002*0.56 [0.37 to 0.83].005* Chronic smokingc3.76 [2.37 to 5.96]<.001*1.68 [0.99 to 2.84].062.01 [1.18 to 3.43].01*General/Cardiovascular Comorbidities Type 2 Diabetes Mellitus1.94 [1.58 to 2.38]<.001*1.53 [1.19 to 1.96]<.001*1.51 [1.18 to 1.94].001* Chronic liver diseased1.34 [0.43 to 4.17].62N/AN/A1.72 [0.54 to 5.49].36 Chronic kidney disease3.38 [2.29 to 4.99]<.001*3.38 [2.18 to 5.21]<.001*3.03 [1.97 to 4.67]<.001* Acute myocardial infarction5.64 [2.89 to 11.0]<.001*5.28 [2.68 to 10.4]<.001*6.26 [3.17 to 12.4]<.001* Heart failure4.07 [2.89 to 5.73]<.001*1.53 [1.03 to 2.25].03*1.90 [1.29 to 2.81].001* Hypertension1.83 [1.53 to 2.19]<.001*1.21 [0.97 to 1.51].101.35 [1.08 to 1.69].01* Ischemic stroke1.90 [1.17 to 3.09].009*1.91 [1.16 to 3.15].011.97 [1.19 to 3.26].008*Breast Cancer Characteristics Hormone replacement therapye0.90 [0.70 to 1.15].39N/AN/A0.98 [0.75 to 1.28].87 Surgical Resection of breast cancerf0.73 [0.62 to 0.86]<.001*0.82 [0.69 to 0.98].030.81 [0.68 to 0.96].02* Presence of metastasisg2.45 [1.96 to 3.06]<.001*2.43 [1.91 to 3.10]<.001*2.50 [1.93 to 3.22]<.001* Radiotherapyh0.87 [0.71 to 1.07].18N/AN/A1.09 [0.86 to 1.38].47 Tamoxifeni0.99 [0.84 to 1.16].89N/AN/A0.88 [0.73 to 1.04].14 Aromatase inhibitorj1.15 [0.96 to 1.39].13N/AN/A1.18 [0.96 to 1.46].11 Targeted Therapyk0.70 [0.29 to 1.70].44N/AN/A1.13 [0.45 to 2.88].79 CDK4/6 Inhibitorl0.05 [0 to 18.9].32N/AN/AN/AN/A Chemotherapym0.47 [0.37 to 0.59]<.001*1.07 [0.82 to 1.41].620.73 [0.56 to 0.97].03* Immune Checkpoint Inhibitorn0.67 [0.17 to 2.70].58N/AN/A0.90 [0.22 to 3.71].88Concurrent anti-inflammatory agents Aspirin/antiplatelet therapyo0.51 [0.41 to 0.62]<.001*0.87 [0.69 to 1.10]0.240.87 [0.69 to 1.09].22 Lipid-lowering therapyp0.31 [0.26 to 0.36]<.001*0.48 [0.39 to 0.57]<.001*0.40 [0.33 to 0.49]<.001* Trimetazidine0.65 [0.47 to 0.92].02*0.48 [0.48 to 0.98].04*0.62 [0.40 to 0.95].03**p≤0.05aHazard ratio (HR) prediction estimates and 95% confidence interval explained by variable of interest as shown by univariable and multivariable Cox proportional hazards regression; Crude Model: unadjusted; Multivariable Model 1: adjusted for potential confounders with P- value ≤0.10 in Crude Model; Multivariable Model 2: adjusted for all potential confounders as defined a prioribSocioeconomic indicator was defined as having received financial support from the comprehensive social security allowance (CSSA) schemecProxied by chronic obstructive pulmonary disease (COPD)dChronic liver disease served as the proximity of the drinking historyeIncluded the use of conjugated estrogen plus medroxyprogesterone acetate (MPA) (Brand names: Prempro, Premphase), Estradiol plus norethisterone acetate (Brand names: Activella, Angeliq), Estradiol plus drospirenone (Brand name: FemHRT), and further search on the hormone replacement therapy procedure (ICD-9 code: 99.24) was usedfICD-9 procedure code of 85.2–85.4.2.4 was employed for all the operations of breast (ICD-9 code of mastectomy: 85.4)gMetastasis to other sits were recorded when the date of metastasis were after the first diagnosis of breast cancerhICD-9 procedure code of 92.2 was employed for all the radiotherapyiIncluded Tamoxifen, Nolvadex and SoltamoxjIncluded Anastrozole (Brand name: Arimidex), Letrozole (Brand name: Femara), and Exemestane (Brand name: Aromasin)kIncluded Trastuzumab (Brand names: Herceptin, Ogivri, Herzuma), Pertuzumab (Brand name: Perjeta), and Ado-trastuzumab emtansine (Brand name: Kadcyla)lIncluded Palbociclib (Brand name: Ibrance), Ribociclib (Brand name: Kisqali), and Abemaciclib (Brand name: Verzenio)mIncluded Anthracycline (Doxorubicin/Adriamycin, Epirubicin/Ellence), Taxanes (Paclitaxel/Taxol, Docetaxel/Taxotere), Cyclophosphamide (Cytoxan), Fluorouracil (5-FU), Carboplatin, and Gemcitabine (Gemzar)nIncluded pembrolizumab (Brand name: Keytruda)oIncluded aspirin, clopidogrel, ticagrelor, and prasugrelpIncluded simvastatin, rosuvastatin, pravastatin, fluvastatin, atorvastatin, ezetimibe, fenofibrate, alirocumab, and evolocumab Predictors for all-cause mortality in patients with pre-existing breast cancers (n=1012)a *p≤0.05 aHazard ratio (HR) prediction estimates and 95% confidence interval explained by variable of interest as shown by univariable and multivariable Cox proportional hazards regression; Crude Model: unadjusted; Multivariable Model 1: adjusted for potential confounders with P- value ≤0.10 in Crude Model; Multivariable Model 2: adjusted for all potential confounders as defined a priori bSocioeconomic indicator was defined as having received financial support from the comprehensive social security allowance (CSSA) scheme cProxied by chronic obstructive pulmonary disease (COPD) dChronic liver disease served as the proximity of the drinking history eIncluded the use of conjugated estrogen plus medroxyprogesterone acetate (MPA) (Brand names: Prempro, Premphase), Estradiol plus norethisterone acetate (Brand names: Activella, Angeliq), Estradiol plus drospirenone (Brand name: FemHRT), and further search on the hormone replacement therapy procedure (ICD-9 code: 99.24) was used fICD-9 procedure code of 85.2–85.4.2.4 was employed for all the operations of breast (ICD-9 code of mastectomy: 85.4) gMetastasis to other sits were recorded when the date of metastasis were after the first diagnosis of breast cancer hICD-9 procedure code of 92.2 was employed for all the radiotherapy iIncluded Tamoxifen, Nolvadex and Soltamox jIncluded Anastrozole (Brand name: Arimidex), Letrozole (Brand name: Femara), and Exemestane (Brand name: Aromasin) kIncluded Trastuzumab (Brand names: Herceptin, Ogivri, Herzuma), Pertuzumab (Brand name: Perjeta), and Ado-trastuzumab emtansine (Brand name: Kadcyla) lIncluded Palbociclib (Brand name: Ibrance), Ribociclib (Brand name: Kisqali), and Abemaciclib (Brand name: Verzenio) mIncluded Anthracycline (Doxorubicin/Adriamycin, Epirubicin/Ellence), Taxanes (Paclitaxel/Taxol, Docetaxel/Taxotere), Cyclophosphamide (Cytoxan), Fluorouracil (5-FU), Carboplatin, and Gemcitabine (Gemzar) nIncluded pembrolizumab (Brand name: Keytruda) oIncluded aspirin, clopidogrel, ticagrelor, and prasugrel pIncluded simvastatin, rosuvastatin, pravastatin, fluvastatin, atorvastatin, ezetimibe, fenofibrate, alirocumab, and evolocumab Table 3Predictors for all-cause mortality in patients with pre-existing or de novo breast cancers (n=2526)aCrude modelMultivariable model 1Multivariable model 2HR [95% CI]P-valueHR [95% CI]P-valueHR [95% CI]P-valueDemographics Age1.08 [1.08 to 1.09]<.001*1.05 [1.04 to 1.05]<.001*1.05 [1.04 to 1.05]<.001* Male0.92 [0.68 to 1.24].59N/AN/A0.93 [0.66 to 1.32].69 Socioeconomic indicatorb0.73 [0.58 to 0.91].005*0.63 [0.50 to 0.79]<.001*0.63 [0.50 to 0.80]<.001* Chronic smokingc2.48 [1.76 to 3.49]<.001*1.66 [1.11 to 2.49].01*1.69 [1.13 to 2.52].01*General/Cardiovascular Comorbidities Type 2 Diabetes Mellitus1.99 [1.75 to 2.27]<.001*1.55 [1.32 to 1.81]<.001*1.54 [1.31 to 1.80]<.001* Chronic liver diseased1.43 [0.68 to 3.00].35N/AN/A1.60 [0.71 to 3.60].26 Chronic kidney disease3.19 [2.41 to 4.23]<.001*3.38 [2.50 to 4.58]<.001*3.40 [2.51 to 4.61]<.001* Acute myocardial infarction1.90 [1.40 to 2.58]<.001*2.49 [1.78 to 3.47]<.001*2.46 [1.76 to 3.44]<.001* Heart failure3.00 [2.50 to 3.48]<.001*1.54 [1.28 to 1.86]<.001*1.54 [1.28 to 1.86]<.001* Hypertension1.68 [1.50 to 1.89]<.001*1.22 [1.06 to 1.41].007*1.22 [1.05 to 1.40].007* Ischemic stroke2.07 [1.58 to 2.69]<.001*1.97 [1.48 to 2.63]<.001*1.98 [1.48 to 2.64]<.001*Breast Cancer Characteristics Hormone replacement therapye0.87 [0.74 to 1.02].090.99 [0.83 to 1.18].890.98 [0.82 to 1.16].79 Surgical Resection of breast cancerf0.62 [0.56 to 0.68]<.001*0.76 [0.68 to 0.85]<.001*0.76 [0.67 to 0.85]<.001* Presence of metastasisg2.73 [2.37 to 3.14]<.001*2.84 [2.43 to 3.32]<.001*2.82 [2.41 to 3.30]<.001* Radiotherapyh0.93 [0.82 to 1.07].31N/AN/A1.07 [0.92 to 1.25].38 Tamoxifeni1.29 [1.16 to 1.43]<.001*1.00 [0.89 to 1.12].981.00 [0.89 to 1.12].99 Aromatase inhibitorj0.79 [0.70 to 0.88]<.001*0.88 [0.78 to 0.99].04*0.87 [0.77 to 0.99].03* Targeted Therapyk0.77 [0.48 to 1.23]<.001*1.45 [0.87 to 2.40].161.44 [0.87 to 2.40].16 CDK4/6 Inhibitorl0.43 [0.11 to 1.72].22N/AN/A0.64 [0.16 to 2.61].54 Chemotherapym0.43 [0.37 to 0.50]<.001*0.72 [0.60 to 0.85]<.001*0.72 [0.60 to 0.85]<.001* Immune Checkpoint Inhibitorn0.70 [0.26 to 1.87].48N/AN/A1.00 [0.37 to 2.70]1.00Concurrent anti-inflammatory agents Aspirin/antiplatelet therapyo0.33 [0.30 to 0.37]<.001*0.89 [0.76 to 1.05]0.170.90 [0.76 to 1.05].18 Lipid-lowering therapyp0.54 [0.47 to 0.62]<.001*0.48 [0.43 to 0.55]<.001*0.48 [0.43 to 0.55]<.001* Trimetazidine0.77 [0.62 to 0.96].02*0.75 [0.60 to 0.94].01*0.76 [0.59 to 0.97].03**p≤0.05aHazard ratio (HR) prediction estimates and 95% confidence interval explained by variable of interest as shown by univariable and multivariable Cox proportional hazards regression; Crude Model: unadjusted; Multivariable Model 1: adjusted for potential confounders with P- value ≤0.10 in Crude Model; Multivariable Model 2: adjusted for all potential confounders as defined a prioribSocioeconomic indicator was defined as having received financial support from the comprehensive social security allowance (CSSA) schemecProxied by chronic obstructive pulmonary disease (COPD)dChronic liver disease served as the proximity of the drinking historyeIncluded the use of conjugated estrogen plus medroxyprogesterone acetate (MPA) (Brand names: Prempro, Premphase), Estradiol plus norethisterone acetate (Brand names: Activella, Angeliq), Estradiol plus drospirenone (Brand name: FemHRT), and further search on the hormone replacement therapy procedure (ICD-9 code: 99.24) was usedfICD-9 procedure code of 85.2–85.4.2.4 was employed for all the operations of breast (ICD-9 code of mastectomy: 85.4)gMetastasis to other sits were recorded when the date of metastasis were after the first diagnosis of breast cancerhICD-9 procedure code of 92.2 was employed for all the radiotherapyiIncluded Tamoxifen, Nolvadex and SoltamoxjIncluded Anastrozole (Brand name: Arimidex), Letrozole (Brand name: Femara), and Exemestane (Brand name: Aromasin)kIncluded Trastuzumab (Brand names: Herceptin, Ogivri, Herzuma), Pertuzumab (Brand name: Perjeta), and Ado-trastuzumab emtansine (Brand name: Kadcyla)lIncluded Palbociclib (Brand name: Ibrance), Ribociclib (Brand name: Kisqali), and Abemaciclib (Brand name: Verzenio)mIncluded Anthracycline (Doxorubicin/Adriamycin, Epirubicin/Ellence), Taxanes (Paclitaxel/Taxol, Docetaxel/Taxotere), Cyclophosphamide (Cytoxan), Fluorouracil (5-FU), Carboplatin, and Gemcitabine (Gemzar)nIncluded pembrolizumab (Brand name: Keytruda)oIncluded aspirin, clopidogrel, ticagrelor, and prasugrelpIncluded simvastatin, rosuvastatin, pravastatin, fluvastatin, atorvastatin, ezetimibe, fenofibrate, alirocumab, and evolocumab Predictors for all-cause mortality in patients with pre-existing or de novo breast cancers (n=2526)a *p≤0.05 aHazard ratio (HR) prediction estimates and 95% confidence interval explained by variable of interest as shown by univariable and multivariable Cox proportional hazards regression; Crude Model: unadjusted; Multivariable Model 1: adjusted for potential confounders with P- value ≤0.10 in Crude Model; Multivariable Model 2: adjusted for all potential confounders as defined a priori bSocioeconomic indicator was defined as having received financial support from the comprehensive social security allowance (CSSA) scheme cProxied by chronic obstructive pulmonary disease (COPD) dChronic liver disease served as the proximity of the drinking history eIncluded the use of conjugated estrogen plus medroxyprogesterone acetate (MPA) (Brand names: Prempro, Premphase), Estradiol plus norethisterone acetate (Brand names: Activella, Angeliq), Estradiol plus drospirenone (Brand name: FemHRT), and further search on the hormone replacement therapy procedure (ICD-9 code: 99.24) was used fICD-9 procedure code of 85.2–85.4.2.4 was employed for all the operations of breast (ICD-9 code of mastectomy: 85.4) gMetastasis to other sits were recorded when the date of metastasis were after the first diagnosis of breast cancer hICD-9 procedure code of 92.2 was employed for all the radiotherapy iIncluded Tamoxifen, Nolvadex and Soltamox jIncluded Anastrozole (Brand name: Arimidex), Letrozole (Brand name: Femara), and Exemestane (Brand name: Aromasin) kIncluded Trastuzumab (Brand names: Herceptin, Ogivri, Herzuma), Pertuzumab (Brand name: Perjeta), and Ado-trastuzumab emtansine (Brand name: Kadcyla) lIncluded Palbociclib (Brand name: Ibrance), Ribociclib (Brand name: Kisqali), and Abemaciclib (Brand name: Verzenio) mIncluded Anthracycline (Doxorubicin/Adriamycin, Epirubicin/Ellence), Taxanes (Paclitaxel/Taxol, Docetaxel/Taxotere), Cyclophosphamide (Cytoxan), Fluorouracil (5-FU), Carboplatin, and Gemcitabine (Gemzar) nIncluded pembrolizumab (Brand name: Keytruda) oIncluded aspirin, clopidogrel, ticagrelor, and prasugrel pIncluded simvastatin, rosuvastatin, pravastatin, fluvastatin, atorvastatin, ezetimibe, fenofibrate, alirocumab, and evolocumab Kaplan-meier curve analyses for all-cause mortality in patients with pre-existing or de novo breast cancers. Trimetazidine group had a longer event-free survival from all-cause mortality compared to control (Mean survival, days [95%CI]; Control, 4,950 [4,839-5,061]; Trimetazidine, 5,531 [5,159-5,902]; Breslow [Generalised Wilcoxon], 8.87, p=.003) Preferential effect of trimetazidine on non-cardiovascular mortality in breast cancer patients Prespecified analyses of non-cardiovascular versus cardiovascular mortality were performed. Trimetazidine users had a longer event-free survival from non-cardiovascular mortality compared to control (Mean survival, days [95%CI]; Control, 5378 [5263-5492]; Trimetazidine, 6034 [5661-6407]; Breslow [Generalised Wilcoxon], 10.6, p=.001, Supplementary Figure S2). Trimetazidine use was associated with a lower risk of non-cardiovascular mortality (HR, 0.69; 95% CI, 0.54–0.89.54.89; P =.004). Conversely, trimetazidine use was associated with worse survival from cardiovascular mortality compared to control (Mean survival, days [95%CI]; Control, 7227 [7138-7315]; Trimetazidine, 7074 [6784-7364]; Breslow [Generalised Wilcoxon], 0.01, p=.91, Supplementary Figure S3). Effect modification by sex There was no effect modification of the primary study findings by sex (P =.59) (Supplementary Figure S4). Discussion To our knowledge, this is the first study in humans that investigated the potential clinical impacts of trimetazidine use on clinical outcomes of breast cancer patients, based on a modulating hypothesis of metabolic reprogramming. In this retrospective study, we found that trimetazidine use was associated with a significant risk reduction of all-cause mortality in pre-existing breast cancer patients (adjusted HR, 0.48; 95% CI, 0.48–0.98.48.98; P =.04). This finding remained statistically robust and were coherent in all breast cancer patients (adjusted HR, 0.75; 95% CI, 0.60–0.94.60.94; P =.01). From the basic scientific perspective, previous laboratory and animal studies have provided support to our theoretical framework, suggesting a potential protective effect of trimetazidine in breast cancer via metabolic reprogramming. The combination of abraxane, a chemotherapy for advanced-stage breast cancer, and high concentrations of trimetazidine was shown to induce more apoptosis than abraxane alone in a cellular model [20]. Studies also suggested that trimetazidine improved the antitumor efficacy of CDK4/6 inhibitor abemaciclib in TNBC, hence delaying tumour growth for 15 days compared to 6 days in abemaciclib alone [21]. As part of the Warburg Effect and related microenvironmental changes, breast cancer cells and their adjacent fibroblasts preferentially undergo glycolysis rather than oxidative phosphorylation even with the presence of oxygen to support cancer cell survival [5, 22]. Whether and how exactly these metabolic reprogramming changes may lead to survival advantage of breast cancer cells remained controversial. In general, trimetazidine shifts the FAO metabolism to glucose metabolism in body cells. In this way, this drug might potentially abolish any comparative evolutionary advantages acquired by breast cancer cells through the conventional Warburg effect. On the other hand, FAO is an essential process to derive ATP in cancer cells [23], such as prostate cancer [24], ovarian cancer [25], TNBC [26]. The inhibition of FAO may therefore directly compromise cancer cell energetics [27], causing cancer cell death without affecting normal cells [28]. Previous studies also showed that FAO inhibition was beneficial in treating [29] and inhibiting metastasis [30] in TNBC. Furthermore, trimetazidine inhibits the glycolytic pathway by enhancing pyruvate dehydrogenase (PDH) and decreasing pyruvate dehydrogenase kinase (PDK) [31], and may reduce breast cancer cell growth by modulating the expression of glycolytic enzymes, including GLUT-1 [15]. Trimetazidine is a second-line therapy for the relief of symptomatic angina, especially in patients with severe IHD who are refractory to first-line treatments. Its role as a second-line agent, coupled with concern regarding drug-induced Parkinsonism, likely contribute to the relatively small sample size in the trimetazidine group. However, post-marketing surveillance data suggested that the trimetazidine-induced Parkinsonism is very low, approximating 0.36 per 100,000 person-years, and typically resolves within 4 months of cessation. Moreover, the complex underlying cardiovascular comorbidities in the trimetazidine group may account for the observed association between trimetazidine use and higher cardiovascular mortality compared to the control group in our study. Nevertheless, the clinical value of trimetazidine therapy over all-cause and non-cardiovascular mortality shown in our study warrants thorough evaluation. Our findings and others support a novel theoretical framework that modulation of metabolic reprogramming may represent a potentially novel therapeutic avenue in breast cancer. The clinical potentials of drug repositioning here illustrated could be immense, as an existing and relatively affordable drug from the drug repertoire designed for other diseases may potentially be employed to impact clinically on breast cancer outcomes. Nevertheless, this is an observational study and such preliminary findings will need to be properly investigated in randomized controlled trial. Limitations Our findings represented important first-in-humans clinical work on mortality outcomes of patients with breast cancer and IHD, but several limitations should be considered when interpreting the results. Firstly, we are unable to establish causality due to the observational nature of the study. Secondly, several important confounders, such as tumor grade, receptor status, or other histopathologic data, were not available. We also did not include data regarding the progression-free survival and response to treatment. Thirdly, while this study was territory-wide, we only included breast cancer patients diagnosed within public hospital system in Hong Kong, which represented 90% of patient catchment in the Hong Kong population. Fourthly, the absence of propensity score matching due to sample size constraints might leave potential residual confounding unaddressed. However, we used multivariable regression models with careful covariate selection to mitigate overfitting risk. Lastly, the lack of detailed mechanistic data on metabolic reprogramming in our clinical sample has deterred us from affirming the causal effects of any specific metabolic pathway. Our discussion on a putative protection by trimetazidine in breast cancer was hypothetical and built upon existing literature. However, direct biological data from our study to support such a postulation were indeed absent. Likewise, causality cannot be established, and mechanistic inferences should be made with caution. Further randomized controlled trials and mechanistic studies are warranted to confirm our postulations further and elucidate the underlying mechanisms. Conclusions Trimetazidine use is observationally associated with a lower risk of all-cause mortality among breast cancer patients with IHD. While causality cannot be established at present, further randomized controlled trials are warranted to confirm our findings and further elucidate the underlying mechanisms. Supplementary Information Supplementary Material 1. Supplementary Material 1.