Advertisement

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Association of hydralazine use with risk of hematologic neoplasms in patients with hypertension: A nationwide population-based cohort study in Taiwan

  • Li-Tzu Wang ,

    Roles Data curation, Formal analysis, Funding acquisition, Investigation, Visualization, Writing – original draft, Writing – review & editing

    ☯ These authors contribute equally as co-first authors to this work.

    Affiliations School of Medical Laboratory Science and Biotechnology, College of Medical Science and Technology, Taipei Medical University, Taipei, Taiwan, Ph.D. Program in Medical Biotechnology, College of Medical Science and Technology, Taipei Medical University, Taipei, Taiwan

    ⨯
  • Wu-Chien Chien ,

    Roles Conceptualization, Data curation, Funding acquisition, Methodology, Project administration, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing

    ☯ These authors contribute equally as co-first authors to this work.

    Affiliations Graduate Institute of Public Health, College of Public Health, National Defense Medical University, Taipei, Taiwan, Department of Medical Research, Tri-Service General Hospital, National Defense Medical University, Taipei, Taiwan, Taiwanese Injury Prevention and Safety Promotion Association, Taipei, Taiwan

    ⨯
  • Kevin Sheng-Kai Ma,

    Roles Data curation, Formal analysis, Investigation, Visualization, Writing – original draft

    Affiliation Center for Global Health, Perelman School of Medicine, University of Pennsylvania, Philadelphia, United States of America

    ⨯
  • Chi-Hsiang Chung,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Visualization, Writing – original draft, Writing – review & editing

    Affiliations Graduate Institute of Public Health, College of Public Health, National Defense Medical University, Taipei, Taiwan, Department of Medical Research, Tri-Service General Hospital, National Defense Medical University, Taipei, Taiwan

    ⨯
  • Yeu-Chin Chen,

    Roles Conceptualization, Methodology, Visualization, Writing – original draft

    Affiliation Division of Hematology and Oncology, Department of Internal Medicine, Tri-Service General Hospital, National Defense Medical University, Taipei, Taiwan

    ⨯
  • Wei-Che Tsai,

    Roles Conceptualization, Methodology, Visualization, Writing – original draft

    Affiliation Division of Cardiology, Department of Internal Medicine, Tri-Service General Hospital, National Defense Medical University, Taipei, Taiwan

    ⨯
  • Bing-Heng Yang

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    rodancer0629@gmail.com

    Affiliations Division of Clinical Pathology, Department of Pathology, Tri-Service General Hospital, National Defense Medical University, Taipei, Taiwan, Graduate Institute of Pathology and Parasitology, College of Medicine, National Defense Medical University, Taipei, Taiwan

    ⨯

Abstract

Background

Onco-hypertension recognizes well-controlled blood pressure as a favorable prognostic factor for survival in patients with hypertension and solid tumors, including hematologic neoplasms. However, it remains unknown whether continuous use of hydralazine—an antihypertensive agent (AHA) with notable anti-neoplastic activity—is associated with a lower risk of hematologic neoplasms compared to other AHAs.

Method and findings

Utilizing Taiwan’s National Health Insurance Research Database, we conducted a 16-year follow-up study (2000–2015) involving 375,107 patients with hypertension treated with an AHA for ≥180 days. The patients with hypertension were divided into two groups based on hydralazine prescription duration: an exposure group (hydralazine ≥180 days; n = 59,786) and a reference group (hydralazine <180 days; n = 239,144) after 1:4 matching for sex, age, and index date with the exposure group. Both groups were well-matched, with a mean age of approximately 60.8 years and 52.19% male. We assess the association between hydralazine use and the risk of hematologic neoplasms using Kaplan–Meier analysis and multivariable Cox proportional hazards regression, with models adjusted for concomitant medications possessing potential anti-neoplastic properties. The 16-year cumulative incidence of hematologic neoplasms was lower in the exposure group (105.58 per 100,000 person-years) than in the reference group (160.33). Accounting for death as competing risk, the exposure group exhibited an adjusted subdistribution hazard ratio (adjusted sHR) of 0.789 (95% confidence interval [0.667,0.913]; P < .001) for hematologic neoplasms compared to the reference group. Subgroup analyses demonstrated that the association with a lower risk was strongest in the longest prescription duration category. For example, for patients with prescription durations of ≥668 days, the adjusted sHR was 0.448 (95% CI [0.366,0.555]; P < .001) for other malignant neoplasms of lymphoid and histiocytic tissue, 0.552 (95% CI [0.453,0.683]; P < .001) for multiple myeloma and immunoproliferative neoplasms, and 0.555 (95% CI [0.457,0.689]; P < .001) for myeloid leukemia. The main limitation was the potential for residual confounding due to the unavailability of lifestyle and laboratory data in the administrative database.

Conclusions

In this study, we observed that long-term hydralazine use in patients with hypertension was associated with a lower, duration-dependent risk of hematologic neoplasms. These findings warrant prospective studies to confirm this association and its potential clinical implications.

Author summary

Why was this study done?

  • High blood pressure is recognized as a risk factor associated with the development of blood cancers.
  • Laboratory studies have shown that hydralazine has biological activities that could counter the mechanisms of blood cancers, such as inhibiting an enzyme called DNA methyltransferase.
  • Despite these promising laboratory findings, there was a significant knowledge gap, as no large-scale, population-based study had investigated whether taking hydralazine long-term was actually associated with a lower risk of these cancers in people.

What did the researchers do and find?

  • We used a national health database from Taiwan to analyze the health records of nearly 300,000 people with high blood pressure over a 16-year period.
  • We compared a group of 59,786 patients who took the drug hydralazine for at least 180 days to a group of 239,144 patients who took other common blood pressure medications.
  • After adjusting for other health factors, we found that the group taking hydralazine long-term had an approximately 21% lower risk of being diagnosed with a blood cancer.

What do these findings mean?

  • Our results suggest a link between the long-term use of hydralazine and a lower risk of developing blood cancers in this population of patients with hypertension.
  • Because this study only observed patients over time and could not account for lifestyle factors or how well patients took their medication, our findings do not prove that hydralazine causes the lower risk.
  • These results highlight the need for future research to confirm the association and understand what it could mean for treating patients with high blood pressure who are at a higher risk for blood cancers.

Citation: Wang L-T, Chien W-C, Ma KS-K, Chung C-H, Chen Y-C, Tsai W-C, et al. (2025) Association of hydralazine use with risk of hematologic neoplasms in patients with hypertension: A nationwide population-based cohort study in Taiwan. PLoS Med 22(12): e1004646. https://doi.org/10.1371/journal.pmed.1004646

Academic Editor: Steven C. Moore, National Cancer Institute, UNITED STATES OF AMERICA

Received: May 8, 2025; Accepted: November 12, 2025; Published: December 4, 2025

Copyright: © 2025 Wang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The data sets used in this study are held by the Taiwan Ministry of Health and Welfare (MOHW). Requests for access to the data sets must be approved by the MOHW. The data sets can be requested by any researcher interested in accessing them. Please visit the website of the National Health Informatics Project of the MOHW (https://dep.mohw.gov.tw/dos/np-2497-113.html) or contact the National Health Insurance Database (nhird@nhri.edu.tw). Address of Taiwan Ministry of Health and Welfare: No. 488, Sec. 6, Zhongxiao E. Rd., Nangang Dist., Taipei City 115204, Taiwan (R.O.C.) Tel: (+886)2-8590-6666. Fax: (+886)2-8590-6000.

Funding: This work was supported by the Tri-Service General Hospital Foundation (TSGH-B-112018 to BHY; TSGH-E-112226 to BHY; TSGH-B-113026 to BHY; TSGH-E-113248 to BHY; TSGH-B-114025 to BHY; TSGH-E-114247 to BHY; TSGH-B-112020 to WCC; TSGH-B-113025 to WCC; TSGH-B-114022 to WCC). The funder’s website is: www.tsgh.ndmutsgh.edu.tw/english. This work was also supported by the National Science and Technology Council, Taiwan (NSTC 114-2314-B-038-089-MY3 to LTW). The funder’s website is: www.nstc.gov.tw/?l=en). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Abbreviations:: adjusted HRs, adjusted hazard ratios; adjusted sHR, adjusted subdistribution hazard ratio; AHA, antihypertensive agent; AML, acute myeloid leukemia; CCI_R, Charlson Comorbidity Index_revised; CHF, congestive heart failure; CVD, cardiovascular disease; DDDs, defined daily doses; DNMT, DNA methyltransferase; HAAEs, hydralazine-associated lupus-like adverse effects; ICD-9-CM, International Classification of Diseases, 9th Revision, Clinical Modification; IHD, ischemic heart disease; LGTD, Longitudinal Generation Tracking Database; MDS, myelodysplastic syndrome; MM, multiple myeloma; NHIRD, Taiwan’s National Health Insurance Research Database; NTD, New Taiwan Dollars; STROBE, Strengthening the Reporting of Observational Studies in Epidemiology

Introduction

Onco-hypertension [1] is an emerging field that recognizes well-controlled blood pressure as a favorable prognostic factor for survival in patients with hypertension and solid tumors or hematologic neoplasms such as high-grade hematological malignancies (HMs) [2]. The link between hypertension and hematologic neoplasms is incompletely understood, especially in high-risk settings, such as after allogeneic hematopoietic cell transplantation, where endothelial injury is a critical pathogenic mechanism driving hypertension [3]. Furthermore, the use of novel targeted agents, including tyrosine kinase inhibitors, can lead to drug-induced hypertension [4], further complicating patient management. This challenge underscores the importance of optimal antihypertensive agents (AHAs) selection, as several agents have not only been associated with a lower risk of specific hematologic neoplasm [5] but also show potential mechanistic action, such as direct anti-proliferative effects [6] or inhibition of pro-tumorigenic β-adrenergic signaling [7]. Moreover, growing bodies of evidence indicate that AHAs can be repurposed for the treatment of hematologic neoplasms by targeting specific biological mechanisms [8–10]. An association between a specific AHA and a lower risk of hematologic neoplasms, relative to other AHAs, would suggest that the drug possesses pleiotropic effects beyond its primary function of blood pressure regulation.

Targeting host susceptibility factors has emerged as a promising strategy for reducing the risk of HMs, especially in high-risk populations [11,12]. This approach is motivated by evidence linking HM development to specific driver genes, such as DNMT3A [13–15] and TET2 [13–15], as well as independent risk factors like hypertension [2,4] and hepatitis B virus infection [2,16]. Accordingly, an ideal agent for reducing the risk of hematologic neoplasms would target multiple pathogenic pathways, including both driver genes and independent risk factors. This highlights a critical gap in the literature: while AHAs possess diverse mechanisms of action and are widely used, their association with the overall risk of hematologic neoplams remains insufficiently investigated.

Certain AHAs have been repurposed for the treatment of specific hematologic neoplasms [8–10]. Hydralazine—an arterial vasodilator and a DNA methyltransferase (DNMT) inhibitor—is among these AHAs and has been repurposed for the treatment of T-cell leukemia [17], cutaneous T-cell lymphoma [18], and myelodysplastic syndrome (MDS) [18,19]. Hydralazine has also been reported to suppress DNMT3a expression [20–22], and DNMT3A mutations are implicated in the development of various hematologic neoplasms [23], including myeloproliferative neoplasms, MDS [24], acute myeloid leukemia (AML) [25–27], and T-cell lymphoma [28]. Furthermore, hydralazine was observed to increase TP53 activity [29], a factor involved in de novo AML [26,27] and lymphomagenesis [30]. Owing to its unique pharmacological profile among AHAs, hydralazine warrants investigation for its potential association with the risk of hematologic neoplasms in patients with hypertension.

According to our review of the literature, the association between AHA use and hematologic neoplasm risk has yet to be evaluated by a large-scale population-based study. Accordingly, to fill this research gap, we used a nationwide database to analyze hematologic neoplasm risk in patients with hypertension receiving hydralazine versus other AHAs.

Methods

Ethics statement

The study protocol was approved by the Institutional Review Board of Tri-Service General Hospital (TSGHIRB No. E202216031). The board waived the requirement for informed consent because of the anonymization of all extracted data.

Data source

This retrospective cohort study employed data extracted from the Longitudinal Generation Tracking Database (LGTD) 2000–2015. The LGTD is a subset of Taiwan’s National Health Insurance Research Database (NHIRD) and encompasses the health records of 1,936,512 patients [31]. Taiwan’s National Health Insurance program provides coverage for ≥99.9% of the country’s 23 million residents [31–33], with the NHIRD serving as the claims database for this program. From the LGTD, we extracted information on the patients’ clinicodemographic characteristics (such as age, sex, and residence area), diagnoses, treatments, and surgical history. Diagnoses were coded using International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) diagnostic codes. Notably, in contrast to unvalidated ICD-10-CM diagnostic codes, the ICD-9-CM diagnostic codes in the NHIRD have been validated to have high sensitivity for hypertension (92.4%) [34] and all cancers (91.5%) [35]. Prior to data extraction, all confidential information, such as medical institutes and patient names, were encrypted to ensure privacy.

Study population and AHA treatments

The study population was selected from LGTD 2000–2015. As a preliminary step to confirm the association between hypertension and hematologic neoplasm in our population, we first identified a cohort of patients with hypertension (ICD-9-CM codes 401–405). This cohort was then matched using propensity scores to patients without hypertension at a 1:4 ratio based on age, sex, and index date (Fig 1). This step confirmed that hypertension was an independent risk factor for hematologic neoplasms in our cohort, providing the rationale for the primary analysis. Medication exposure was standardized by converting all prescription doses into defined daily doses (DDDs), as specified by the World Health Organization Collaborating Centre for Drug Statistics Methodology. The DDD is the assumed average maintenance dose per day for a drug used for its primary indication in adults. For hydralazine, the DDD is 0.1 g per day. The cumulative exposure for each patient was calculated by dividing the total prescribed dose of hydralazine recorded in the database by its DDD. This method allowed for a standardized assessment of exposure duration and for classifying patients into an exposure group (≥180 days of cumulative use) and a reference group (<180 days of cumulative use). To enhance comparability, we randomly selected a subset of patients from the reference group for propensity score matching with those in the exposure group in terms of age, sex, and index date at a 4:1 ratio. Considering that the diagnostic codes for hypertension may not have been recorded for some patients with hypertension receiving regular AHA treatment, which would have resulted in an underestimation of the study population, we also included AHA-treated patients who had received a hypertension diagnosis within the 2 years prior to the index date. The cumulative incidence of hematologic neoplasm was estimated using Kaplan–Meier curves. We employed the 2020 International Society of Hypertension global hypertension practice guidelines [36–38] for the selection of AHAs, namely the A/C/D classes of AHAs: A (angiotensin-converting enzyme inhibitors or angiotensin receptor blockers: quinapril hydrochloride, lisinopril, fosinopril sodium, enalapril maleate, perindopril, captopril, benazepril hydrochloride, and ramipril), C (calcium channel blockers: nifedipine, felodipine, nicardipine, amlodipine besylate, verapamil hydrochloride, and diltiazem hydrochloride), and D (thiazide-like diuretics: chlorthalidone, chlorothiazide, indapamide, and metolazone), along with hydralazine. Other AHAs (spironolactone, α-blockers, and β-blockers) were only considered for specific indications (e.g., hyperkalemia, atrial fibrillation, heart failure, angina, and younger women who were pregnant or were planning pregnancy) [36] not for general use, and thus we did not include them in this study. Only patients who received any (single or combination therapy) of the aforementioned A/C/D classes of AHAs and hydralazine successively for ≥180 days were included. After confirming patient eligibility, we calculated person-time for exposed and unexposed patients. For exposed patients, person-time was calculated from the date they accrued ≥180 days of hydralazine exposure. For unexposed patients, person-time was calculated from the date they accrued ≥180 days of exposure to A, C, or D classes of AHAs. Person-time was measured from the start of follow-up until the date of hematologic neoplasm development, date of death, or the end of follow-up, whichever occurred first. Regarding the exclusion criteria, participants who were aged <20 years, were lost to follow-up, received a diagnosis of hematologic neoplasm before the index date, or had missing demographic information excluded. To assess potential selection bias, we compared baseline demographic and clinical characteristics between the included patients (exposure group, n = 59,786) and those initially excluded from the population of patients with hypertension (n = 1,026) (Table A in S1 File).

thumbnail
Download:
Fig 1. Population-based analysis of hematologic neoplasm development in patients without hypertension and patients with hypertension receiving regular antihypertensive agents.

https://doi.org/10.1371/journal.pmed.1004646.g001

Covariates and comorbidities

We employed sex, age (20–29, 30–39, 40–49, 50–59, or ≥60 years), season of index date, residence area, urbanization level (1: ≥1,250,000 people; 2: 500,000–1,249,999 people; 3: 150,000–499,999 people; or 4: <149,999 people), health insurance premiums, and hospital level (medical center, regional hospital, or local hospital) as covariates. Health insurance premiums, denominated in New Taiwan Dollars (NTD), are calculated based on income levels and serve as a reliable proxy for the patient’s socioeconomic status within Taiwan’s National Health Insurance system. In this study, premiums were categorized into three groups as NTD per month: <18,000, 18,000–34,999, and ≥35,000 (1 NTD = ~0.03 USD).

We also adjusted potential confounders, such as the comorbidities [39] or concomitant medications [8,40] in which previous studies have reported direct or indirect associations with hematologic neoplasm development (Table B in S1 File). Additionally, the Charlson Comorbidity Index_revised (CCI_R) was used to evaluate the overall extent of the comorbidity-associated hematologic neoplasm risk.

Outcome measure

The primary outcome was the occurrence of any hematologic neoplasm event in a patient. A hematologic neoplasm event was identified on the basis of the presence of any of the following ICD-9-CM codes: (1) 200 (lymphosarcoma and reticulosarcoma); (2) 201 (Hodgkin’s disease); (3) 202 (other malignant neoplasms of lymphoid and histiocytic tissue); (4) 203 [multiple myeloma (MM) and immunoproliferative neoplasms]; (5) 204 (lymphoid leukemia); (6) 205 (myeloid leukemia); (7) 206 (monocytic leukemia); (8) 207 (other specified leukemia); (9) 208 (leukemia of unspecified cell type); (10) 238.4, 238.5, 238.6, 238.71–238.76, 238.79, or 289.83 (neoplasm of uncertain behavior); (11) 238.72–238.75 (MDS); (12) 273.1–273.3 or 273.8–273.9 (paraproteinemia); and (13) 289.0 or 289.6 (other polycythemia; Table B in S1 File).

The onset and long-term progression of hematologic neoplasms were analyzed using two models (Table C in S1 File). The first (first-event) model was based on initial diagnosis to assess the risk of developing hematologic neoplasm for the first time. Given that the clinical course of hematologic neoplasms over 16 years can be complex, the second (multiple-event) model included all hematologic neoplasm events from each patient to evaluate cumulative disease burden. The use of both models enabled us to conduct a comprehensive analysis: the first-event model could capture the initial risk of hematologic neoplasm, whereas the multiple-event model could capture the cumulative burden and progression of the disease, thereby enhancing the understanding of both hematologic neoplasm onset and long-term outcomes. Because patients could receive multiple hematologic neoplasm diagnoses during the follow-up period, an overall adjusted subdistribution hazard ratio (adjusted sHR) could not be calculated in the multiple-event model. The association between the incidence of hematologic neoplasms and mortality in the exposure group was evaluated by calculating hematologic neoplasm-related and all-cause mortality. Participants with any diagnosis of hematologic neoplasm on the date of their mortality were considered as having hematologic neoplasm-related mortality.

Statistical analysis

Intergroup comparisons of categorical variables were performed using a chi-squared test or Fisher’s exact test, depending on whether the proportion of all categorical outcomes was >5% or any proportion was <5%, respectively. Continuous variables were compared using a t test or one-way analysis of variance with Scheffe’s post hoc test. To assess the cumulative incidence of hematologic neoplasms, the log-rank test was employed, and the results were visualized using Kaplan–Meier curves. Associations with hematologic neoplasms were determined using multivariable Cox regression analyses, with results presented as adjusted hazard ratios (adjusted HRs) along with 95% confidence intervals. Statistical significance was set at a two-tailed P value of <.001. This stringent threshold was chosen to minimize the risk of false positives (type I error) owing to the large sample size and multiple statistical comparisons. Schoenfeld’s global test was conducted using STATA 9.0 to evaluate the proportionality assumption of covariates and comorbidities [41]. To account for the potential impact of disproportionate subgroup distributions on the overall results, we conducted a leave-one-out analysis by excluding cases from any specific subgroup that constituted >30% of the study population. Moreover, a sensitivity analysis was performed by excluding patients who received a hematologic neoplasm diagnosis within the first few years of tracking. For estimating the competing risk of mortality, Fine and Gray’s competing risk model was constructed, with all-cause mortality serving as a covariate [42]. The hematologic neoplasm risk was estimated using two models: adjusted HRs (based on Cox regression), and adjusted sHRs (based on Fine and Gray’s competing risk models) (Table 2); the corresponding unadjusted (crude) hazard ratios are provided in the Table K–TableQ in S1 File. All analyses were conducted using SPSS (version 22.0; IBM, Armonk NY, USA). This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (S1 Checklist).

thumbnail
Download:
Table 1. Baseline characteristics of patients with hypertension by prescription duration of hydralazine, 2000–2015.

https://doi.org/10.1371/journal.pmed.1004646.t001

thumbnail
Download:
Table 2. Multivariable risk regression analysis of hematologic neoplasm development in patients with hypertension in competing risk model*.

https://doi.org/10.1371/journal.pmed.1004646.t002

Results

Patient selection and characterization

The study population was established through two 1:4 matching procedures (Fig 1). First, we identified a primary cohort of 375,107 patients with hypertension and matched them to 1,500,428 patients without hypertension (selected from a pool of 1,560,379 individuals). Second, within the cohort of patients with hypertension, we defined an exposure group (n = 59,786) and a reference pool (n = 315,321). We then matched the exposure group 1:4 to this reference pool, yielding a final reference group of 239,144. The 16-year cumulative hematologic neoplasm incidence in the exposure group was significantly lower than that in the reference group (105.58 versus 160.33 per 100,000 person-years).

To check for selection bias, we compared baseline demographic and clinical characteristics between the exposure group (n = 59,786) and excluded patients (n = 1,026) (Table A in S1 File). The excluded individuals were significantly younger (mean age 52.18 versus 60.82 years, P < .001) and had a significantly higher proportion of men (66.08% versus 52.19%, P <.001); in addition, the excluded individuals had a significantly higher comorbidity burden, evidenced by higher rates of CHF (5.46% versus 1.33%, P <.001) and higher mean CCI_R scores (1.05 versus 0.82, P <.001).

Because of the propensity score matching process, age (60.79 versus 60.82 years in the reference group and exposure group), sex (male-to-female ratio: approximately 1.09 in both), and index date (proportions in the 4 seasons) were comparable between the 2 groups. More than half of the patients in both groups were older than 60 years. Regarding residence area, approximately one-third of the patients resided in northern Taiwan, and approximately 70% of them resided in high urbanization levels of the city (1 and 2). The exposure group had a significantly higher comorbidity burden, with higher rates of congestive heart failure (CHF), ischemic heart disease (IHD), malignant neoplasm of the kidney/renal pelvis, and acute glomerulonephritis/nephrotic syndrome, as well as a higher overall CCI_R score. Furthermore, a larger proportion of patients in the exposure group received aspirin, celecoxib, thalidomide, valproate, auranofin, ivermectin, curcumin, or axitinib (all P < .001), while a smaller proportion received mebendazole (P < .001) and statins (P = .027) (Table 1).

Risk factors associated with hematologic neoplasms

After adjustment of potential confounders, hypertension was associated with a higher risk of hematologic neoplasms (adjusted sHR = 1.483, 95% confidence interval [1.397,1.654]; P < .001; Table D in S1 File). Patients with IHD, vascular insufficiency of the intestine, hepatitis B virus infection with or without hepatic coma, malignant neoplasm of the kidney, acute glomerulonephritis or nephrotic syndrome, or a higher CCI_R score were at higher risk of hematologic neoplasms (Table 2). We also examined the associations for other medications known to have anti-hematological neoplasm properties. Mebendazole, which has recognized antileukemia activity [43], was associated with a lower risk of hematologic neoplasms (adjusted sHR = 0.696, 95% CI [0.286,0.957]; P = .003), although this association did not meet our pre-specified significance threshold. Similarly, no statistically significant associations were observed for other medications investigated: itraconazole [44] and metformin [45] were linked to a lower risk, whereas valproate [46] was linked to a higher risk. However, a higher risk of hematologic neoplasms was associated with the use of aspirin, celecoxib, statins, and thalidomide—with the associations for celecoxib (adjusted sHR = 1.608, 95% CI [1.163,1.933]; P < .001) and thalidomide (adjusted sHR = 1.846, 95% CI [1.194,2.466]; P < .001) meeting our pre-specified significance threshold—all of which were previously documented to play a therapeutic role in various types of hematologic neoplasms [47].

Association of hematologic neoplasm incidence stratified by hematologic neoplasm subgroup and duration of hydralazine prescription

After accounting for the competing risk of mortality, we observed a duration-dependent association between hydralazine use and a lower risk of overall hematologic neoplasms (Table 3). In the exposure group, patients were categorized into three subgroups based on prescription duration: 180–350 days, 351–667 days, and ≥668 days. As detailed in Table 3, the adjusted HRs were 0.884 (95% confidence interval [0.743,1.098]; P = .189) for the 180–350 days subgroup, 0.728 (95% CI [0.598,0.904]; P < .001) for the 351–667 days subgroup, and 0.646 (95% CI [0.531,0.803]; P < .001) for the ≥ 668 days subgroup.

thumbnail
Download:
Table 3. Adjusted hazard ratio of hematologic neoplasm development for overall risk and key subgroups, stratified by prescription duration of hydralazine.

https://doi.org/10.1371/journal.pmed.1004646.t003

A statistically significant, inverse duration-response association was observed between hydralazine use and hematologic neoplasm risk across several subgroups. This association was most pronounced for other malignant neoplasms of lymphoid and histiocytic tissue, MM and immunoproliferative neoplasms, and other polycythemia; these key findings were presented in Table 3. The detailed analyses for all other hematologic neoplasm subgroups were provided in Table E in S1 File.

This association suggested a duration-dependent pattern, with a greater reduction in risk observed with longer prescription durations, a finding that was consistent across both models (Table C in S1 File). The median duration of hydralazine prescription was 9.18 years (Table G in S1 File), and the corresponding data exhibited an approximately normal distribution. Notably, hydralazine was associated with a lower risk of subsequent development of leukemia of unspecified cell type, as shown in the multiple-event model (Table C in S1 File).

Leave-one-out analysis stratified by hematologic neoplasm subgroup and duration of hydralazine prescription

To confirm that the observed inverse association was not disproportionately driven by the “other polycythemia” subgroup, which constituted 38% of all hematologic neoplasm cases, we performed a leave-one-out sensitivity analysis. Accordingly, an additional subgroup analysis was performed to determine the composition of other polycythemia. As displayed in Table I in S1 File, 2,025 individuals received a diagnosis of other polycythemia over the 16-year follow-up period, of whom 2,005 had secondary polycythemia (90.9%) and 20 had familial polycythemia (0.99%). The leave-one-out analysis (Table F in S1 File) demonstrated the statistical robustness of the finding, as the inverse association between hydralazine use and hematologic neoplasms persisted after excluding cases of secondary polycythemia.

Association of long-term hydralazine use with hematologic neoplasm incidence

Long-term hydralazine use was associated with a longer time to hematologic neoplasm diagnosis. The median interval from the index date to the first hematologic neoplasm diagnosis was significantly longer in the exposure group than in the reference group (7.33 versus 7.05 years; P < .001; Table H in S1 File). After the 16-year follow-up, hematologic neoplasm had been diagnosed in 757 patients (1.27%) in the exposure group and 4,544 (1.90%) in the reference group (Table I in S1 File). Compared with that in the reference group, the adjusted HRs of overall hematologic neoplasm incidence in the exposure group after adjustment for covariates and comorbidities were 0.762 (95% confidence interval [0.653,0.897]; P < .001; Table 3). Additionally, the exposure group exhibited a significantly lower cumulative incidence of overall hematologic neoplasm (log-rank P value in the 16th year being <.001; Fig 2A) than did the reference group. This pattern was also seen in the analysis of cumulative incidence, which was significantly lower in the exposure group for several hematologic neoplasm subgroups, including MM and immunoproliferative neoplasm (Fig 2B), myeloid leukemia (Fig 2C), neoplasm of uncertain behavior (Fig 2D), and other polycythemia (Fig 2E) (all log-rank P < .001).

thumbnail
Download:
Fig 2. Association of long-term hydralazine use with the incidence of hematologic neoplasms in patients with hypertension, 2000–2015.

The cumulative incidence of hematologic neoplasms in the reference group (hydralazine exposure <180 days) is indicated by the orange line, and that in the exposure group (hydralazine exposure ≥180 days) is indicated by the blue line. (A) Cumulative incidence of overall hematologic neoplasm was significantly lower. (B–E) Cumulative incidence stratified by hematologic neoplasm subgroups was significantly lower in the following subgroups: (B) multiple myeloma and immunoproliferative neoplasm, (C) myeloid leukemia(), (D) neoplasms of uncertain behavior, and (E) other polycythemia.

https://doi.org/10.1371/journal.pmed.1004646.g002

Sensitivity analysis for the hematologic neoplasm incidence

To minimize potential selection bias arising from the inclusion of patients with ongoing hematologic neoplasm in the early stages of follow-up, a sensitivity analysis was performed by excluding patients diagnosed within the first year or the first 5 years. Table 4 presents the results of this analysis. After excluding patients diagnosed within the first year, the association for the overall ≥180 days group was no longer statistically significant (adjusted sHR = 0.787, 95% confidence interval [0.646,0.980]; P = .03). Similarly, after excluding the first five years, the association for this group did not meet our pre-specified threshold (adjusted sHR = 0.798, 95% CI [0.659,0.950]; P = .001). However, the duration-dependent pattern remained robust; the association for the longest-duration subgroup (≥668 days) remained statistically significant in both the 1-year exclusion (adjusted sHR = 0.674, 95% CI [0.555,0.838]; P < .001) and 5-year exclusion (adjusted sHR = 0.684, 95% CI [0.565,0.858]; P < .001) analyses.

thumbnail
Download:
Table 4. Sensitivity analysis for factors of hematologic neoplasm development by using Cox regression in competing risk model.

https://doi.org/10.1371/journal.pmed.1004646.t004

Mortality analysis

The mortality analysis showed that there was no significant difference in hematologic neoplasm-related mortality between the exposure group and reference group (adjusted HR = 0.884, 95% confidence interval [0.632,1.238]; P = .265, Table J in S1 File). Similarly, all-cause mortality was not significantly different between the two groups (adjusted HR = 1.075, 95% CI [0.768,1.506]; P = .536). These results indicate that hydralazine use was associated with a lower risk of developing hematologic neoplasm but did not significantly alter long-term survival outcomes. The absence of an observed difference in mortality, despite the lower risk of hematologic neoplasms in the exposure group, might be attributable to the higher baseline comorbidity burden in these patients. This higher burden could have masked any potential association between a lower risk of hematologic neoplasms and survival.

Discussion

This nationwide retrospective cohort study investigated the association between the use of AHAs and the risk of developing hematologic neoplasms in patients with hypertension. Our findings indicate that the exposure group had an approximately 21% lower risk of overall hematologic neoplasm than did the reference group, particularly for other malignant neoplasms of lymphoid and histiocytic tissue (43% lower), MM and immunoproliferative neoplasms (38% lower), and myeloid leukemia (29% lower). Furthermore, multivariable analysis revealed a duration-dependent inverse association between hydralazine use and the risk of hematologic neoplasms in patients with hypertension.

Cardiovascular disease (CVD) and cancers are the top 2 leading causes of mortality worldwide [48]. Notably, hypertension is not only the leading cause of CVD [49] but also a risk factor for various types of cancer; hence, determining an optimal strategy for hypertension management could reduce global mortality. A nationwide cohort study [2] and the current study both indicated a relatively high risk of hematologic neoplasm development in patients with hypertension, signifying that hypertension is associated not only with CVD and solid tumors but also with hematologic neoplasms. Although the association between the use of AHAs and the risk of certain solid tumors has been explored—with their involvement in biological functions such as reducing inflammation and angiogenesis being documented in preclinical studies [9]—the specific mechanisms underlying a potential association with hematologic neoplasms remain largely unclear, and further clinical investigation is needed.

In the present study, long-term hydralazine use was associated with a significantly lower risk of hematologic neoplasms across several subgroups; this finding suggests the association may be linked to several biological regulatory mechanisms. In addition to its antihypertensive action, hydralazine targets DNMT [50], enhances P53 function [51], and participates in other crucial anti-hematologic neoplasm signaling pathways. Hydralazine has been demonstrated to reduce the viability of monocytic leukemia cells [52] and to counteract chemoresistance in chronic myeloid leukemia [53], a clonal disorder associated with the Philadelphia chromosome, which results from the t(9;22) translocation and carries the BCR-ABL fusion gene that encodes the oncogenic BCR-ABL protein. This chimeric protein leads to the aberrant activation of several signaling pathways, including the PI3K-AKT pathway. Hydralazine has also been suggested to inhibit AKT activation (as observed in an animal sepsis model [54]) and to reduce cleaved caspase-3 and caspase-9 levels (as shown in a rat model of cardiac injury [55]), and has been demonstrated to promote caspase-dependent apoptotic cell death in human leukemic T cells [17]. Moreover, hydralazine inhibits glutamic-oxaloacetic transaminase 1 (a finding from an in vitro screening assay [56]), a prognostic marker of AML [57], indicating its potential role in reducing AML risk. Hydralazine also inhibits angiogenesis (a finding from both in vitro and animal studies [58]) by suppressing vascular endothelial growth factor and basic fibroblast growth factor signaling, both of which are correlated with the clinicopathological features of myeloproliferative neoplasms [59] and MM [60]. Overall, in addition to its role in managing hypertension, hydralazine use was associated with a lower risk of hematologic neoplasms. This association may be explained by its activity in several biological pathways.

Conversely, studies exploring the association between the use of NSAIDs—such as aspirin [61], celecoxib [62], and thalidomide [63]—and the risk of hematologic neoplasms have reported inconsistent findings. For example, some studies have described an association with a decreased risk of certain hematologic neoplasms, others have found no association [64], while some have observed an association with an increased risk [65]. Most NSAIDs have been reported to interfere with the therapeutic action of AHAs [66], an interaction that may be associated with increased blood pressure. This proposed mechanism may help explain the association with a higher risk of hematologic neoplasms that was observed for celecoxib and thalidomide in our 16-year follow-up cohort study. While the association between aspirin, a unique NSAID, and blood pressure remains controversial, some cohort studies have reported that aspirin use is linked to an increased risk of developing hypertension [67]. It remains unclear whether aspirin increases or decreases cancer risk. A meta-analysis of cohort studies [68] revealed that when used at low doses, aspirin can reduce the risk of colorectal cancer, but at high doses, it can increase the risks of lung cancer and prostate cancer. Additionally, aspirin was reported to accelerate the progression of both solid cancers and HMs in older adults [69]. The discrepancies in the reported association between aspirin use and cancer risk may be attributable to methodological heterogeneity across studies, such as variations in dosage, study populations, or the specific cancer types investigated. In the present cohort study, more than 50% of the enrolled patients were aged ≥60 years (Table 1) and exhibited a higher incidence of hypertension and lower immune surveillance, which resulted in an increased hematologic neoplasm risk. Therefore, considering the elevated baseline risk in this older population with hypertension, our findings—which include an observed association between certain NSAIDs and an increased risk of hematologic neoplasms—do not support a potential risk-reducing role for these agents in this context.

Axitinib, a tyrosine kinase inhibitor with antileukemic activity [70], has been associated with the induction and exacerbation of hypertension [71], which may contribute to a higher risk of hematologic neoplasms; this finding was consistent with our study findings (adjusted HR = 1.303, 95% confidence interval [1.000,1.525]; P = .005). Notably, the broad-spectrum anthelmintic medication mebendazole was associated with a lower risk of hematologic neoplasm development in this study. Mebendazole has previously been reported to inhibit the growth of various AML cell lines and mononuclear cells derived from the bone marrow of patients with AML in vitro. This inhibitory effect is thought to be mediated by the downregulation of Akt and Erk signaling pathways [43]. However, to our knowledge, no cohort study has reported an association between mebendazole use and the risk of leukemia in patients with hypertension.

Although the association between hypertension and a higher risk of several cancers has been established, evidence regarding the effect of AHAs on cancer risk is inconsistent [72]. The potential anticancer efficacy of AHAs may be diminished by the interferences resulting from simultaneous multidrug interactions. Furthermore, cohort studies with insufficient control for confounders are prone to spurious associations that can mask, or even invert, the true relationship between an AHA and cancer risk.

A critical consideration for these findings, however, is the known safety profile of long-term hydralazine use. The potential for dose-dependent adverse effects, such as hydralazine-associated lupus-like adverse effects (HAAEs), raises major concerns. According to a previous cohort study involving 36,349 patients with hypertension [2], a daily dose of <34 mg was associated with a significantly lower risk of overall hematologic neoplasm (adjusted HR = 0.791, 95% confidence interval [0.578,0.927]; P < .001) when compared with hydralazine non-users. However, no case of HAAEs was reported in patients receiving a daily dose of 50 mg [73], suggesting that the dosage (<34 mg per day) associated with a lower risk of hematologic neoplasm was considerably below the dosage at which HAAEs have been observed. Although slow acetylators are generally considered more susceptible to HAAEs [74], HAAEs have rarely been reported among patients receiving a daily hydralazine dose of <50 mg, regardless of their acetylator status.

Our study has several strengths, including its large sample size and its use of verified information for evaluating long-term hydralazine-associated hematologic neoplasm risks. However, this study has some limitations that should be considered. First, potential misinformation may have arisen from errors in the NHIRD. Second, the LGTD lacks data on key behavioral and socioeconomic confounders. Information on lifestyle factors such as smoking, alcohol consumption, and physical activity, as well as formal socioeconomic status indicators beyond insurance premiums, was unavailable for adjustment. Third, the relationship between hematologic neoplasm severity and hypertension was not evaluated. Fourth, the lack of available laboratory data limited our ability to identify the potential mechanisms underlying the observed association between hydralazine use and hematologic neoplasm development. Fifth, the study did not include several AHAs (spironolactone, α-blockers, and β-blockers) for comparison with hydralazine, potentially introducing bias. However, the antineoplastic efficacy of these agents for hematologic neoplasms is not yet reported, except for prazosin [75]. Finally, we could not directly contact patients to verify their use of hydralazine and medication compliance due to their anonymous identities. Although some patients with hypertension may have had poor medication adherence, our consideration of a prescription period of ≥180 days may have minimized this potential bias. Accordingly, the observed association between hydralazine use and a lower risk of hematologic neoplasm development remained, despite the possible underestimation of the actual dosage of hydralazine. This finding suggests that the association between hydralazine and a lower risk of hematologic neoplasm may be relevant in real-world clinical practice.

Despite these limitations, we believe that our retrospective study provides real-world evidence and valuable insights into the association between the use of antihypertensive hydralazine and the risk of hematologic neoplasms in patients with hypertension. Although the observational design cannot establish causality, the proposed association is supported by several credible findings, including a duration-dependent relationship and consistent results across multiple analyses. Clinically, the combination of hydralazine and valproate has shown activity in the treatment of MDS [19] and cutaneous T-cell lymphoma [76]. This clinical observation provides a parallel to the association found in our study between hydralazine use and a lower risk of hematologic neoplasms. Furthermore, to check for selection bias, we compared baseline demographic and clinical characteristics between exposure group and those excluded during initial screening. The excluded individuals were clinically distinct but had been removed prior to propensity score matching. Specifically, our leave-one-out analysis indicated that the association between hydralazine use and a lower risk of hematologic neoplasm persisted even after excluding cases from a major hematologic neoplasm subgroup, supporting the robustness of the observed association for overall hematologic neoplasms. These methodological approaches ensured that the included cohorts were well-balanced, thereby minimizing selection bias. In conclusion, our results highlight hydralazine as a compelling candidate for drug repurposing to address the risk of hematologic neoplasms. Such a strategy is advantageous because it may circumvent the protracted timelines and substantial costs inherent in novel drug development.

It is crucial to consider our findings within the clinical context of hydralazine use. As a later-line AHA, hydralazine is often prescribed to patients with more severe or refractory hypertension, heart failure, or chronic kidney disease. Our baseline data reflected this reality, as the exposure group had a significantly higher burden of cardiovascular and renal comorbidities, including CHF, IHD, and glomerulonephritis, as well as higher CCI_R scores (Table 1). This confounding by indication would typically bias the results towards an increased risk of adverse outcomes in the hydralazine group. Therefore, the observation of a significantly lower risk of hematologic neoplasms in the group with a higher comorbidity burden—an association that persisted after multivariable adjustment—strengthens the robustness of our findings. For patients with an existing indication, hydralazine use may be associated with a lower risk of hematologic neoplasm. Nonetheless, prospective studies are warranted to further investigate this association and its potential clinical implications.

In conclusion, our findings indicate that the use of hydralazine in patients with hypertension is associated with a significantly reduced risk of hematologic neoplasms. This association suggests that for patients with hypertension, particularly those with multiple susceptibility factors for hematologic neoplasms, hydralazine use may be linked to a lower incidence of these neoplasms, warranting further prospective studies to investigate this relationship.

Supporting information

S1 Checklist. This checklist is provided in accordance with the STROBE statement, available from https://www.strobe-statement.org/.

https://doi.org/10.1371/journal.pmed.1004646.s001

(DOCX)

S1 File.

Table A. Baseline characteristics of the hydralazine cohort compared to patients with hypertension excluded during initial screening. Table B. ICD-9-CM coding and definition. Table C. Comparison of the adjusted subdistribution hazard ratio of hematologic neoplasms according to subgroup stratified by prescription duration of hydralazine in first-event and multiple-event models in a competing risk model. Table D. Multivariable risk regression analysis of hematologic neoplasm development in patients without/with hypertension in competing risk model. Table E. Adjusted hazard ratio for remaining hematologic neoplasm subgroups, stratified by prescription duration of hydralazine. Table F. Leave-one-out analysis for comparison of adjusted hazard ratio of hematologic neoplasms according to subgroup stratified by prescription duration of hydralazine in first-event and multiple-event models in a competing risk model. Table G. Tracking years in patients with hypertension by prescription duration of hydralazine. Table H. Tracking years from initiating hydralazine prescription to having hematologic neoplasms in patients with hypertension. Table I. Endpoint characteristics of patients with hypertension by prescription duration of hydralazine, 2000–2015. Table J. Mortality analysis of patients with hypertension by prescription duration of hydralazine, 2000–2015. Table K. Unadjusted (crude) hazard ratios for risk factors associated with hematologic neoplasm development. Table L. Unadjusted (crude) hazard ratios for hematologic neoplasm development, stratified by prescription duration of hydralazine. Table M. Unadjusted (crude) hazard ratios for sensitivity analysis of hematologic neoplasm development. Table N. Unadjusted (crude) subdistribution hazard ratios for first-event and multiple-event models. Table O. Unadjusted (crude) hazard ratios for hematologic neoplasm risk associated with hypertension. Table P. Unadjusted (crude) subdistribution hazard ratios for leave-one-out sensitivity analysis. Table Q. Unadjusted (crude) hazard ratios for mortality analysis.

https://doi.org/10.1371/journal.pmed.1004646.s002

(DOCX)

Acknowledgments

The authors thank Wallace Academic Editing (www.editing.tw/) and Editage (www.editage.com.tw) for editing the manuscript.

Disclaimer

This study is based in part on data from the National Health Insurance Research Database provided by the Bureau of National Health Insurance, Department of Health, and managed by National Health Research Institutes.

References

  1. 1. Sahni G. Onco-hypertension: changing paradigm of treating hypertension in patients with cancer. J Clin Oncol. 2023;41(5):958–63. pmid:36332165
  2. 2. Yang B-H, Lin W-Z, Chiang Y-T, Chen Y-C, Chung C-H, Chien W-C, et al. Epigenetics-associated risk reduction of hematologic neoplasms in a nationwide cohort study: the chemopreventive and therapeutic efficacy of hydralazine. Front Oncol. 2022;12:809014. pmid:35186746
  3. 3. Evangelidis P, Gavriilaki E, Tsakiris DA. Thrombotic complications after hematopoietic stem cell transplantation and other cellular therapies. Thrombosis Update. 2024;16:100186.
  4. 4. Gavriilaki E, Gkaliagkousi E, Grigoriadis S, Anyfanti P, Douma S, Anagnostopoulos A. Hypertension in hematologic malignancies and hematopoietic cell transplantation: an emerging issue with the introduction of novel treatments. Blood Rev. 2019;35:51–8. pmid:30898309
  5. 5. Huang W, Sundquist K, Sundquist J, Ji J. Use of dipyridamole is associated with lower risk of lymphoid neoplasms: a propensity score-matched cohort study. Br J Haematol. 2022;196(3):690–9. pmid:34553368
  6. 6. Stanojkovic TP, Zizak Z, Mihailovic-Stanojevic N, Petrovic T, Juranic Z. Inhibition of proliferation on some neoplastic cell lines-act of carvedilol and captopril. J Exp Clin Cancer Res. 2005;24(3):387–95. pmid:16270525
  7. 7. Lamkin DM, Sloan EK, Patel AJ, Chiang BS, Pimentel MA, Ma JCY, et al. Chronic stress enhances progression of acute lymphoblastic leukemia via β-adrenergic signaling. Brain Behav Immun. 2012;26(4):635–41. pmid:22306453
  8. 8. Allegra A, Imbesi C, Bitto A, Ettari R. Drug repositioning for the treatment of hematologic disease: limits, challenges and future perspectives. Curr Med Chem. 2021;28(11):2195–217. pmid:33138750
  9. 9. Regulska K, Regulski M, Karolak B, Michalak M, Murias M, Stanisz B. Beyond the boundaries of cardiology: still untapped anticancer properties of the cardiovascular system-related drugs. Pharmacol Res. 2019;147:104326. pmid:31340189
  10. 10. Regulska K, Regulski M, Karolak B, Murias M, Stanisz B. Can cardiovascular drugs support cancer treatment? The rationale for drug repurposing. Drug Discov Today. 2019;24(4):1059–65. pmid:30878563
  11. 11. Haring B, Wissel S, Manson JE. Somatic mutations and clonal hematopoiesis as drivers of age-related cardiovascular risk. Curr Cardiol Rep. 2022;24(8):1049–58. pmid:35657494
  12. 12. Ye B, Sheng Y, Zhang M, Hu Y, Huang H. Early detection and intervention of clonal hematopoiesis for preventing hematological malignancies. Cancer Lett. 2022;538:215691. pmid:35469959
  13. 13. Asada S, Kitamura T. Clonal hematopoiesis and associated diseases: a review of recent findings. Cancer Sci. 2021;112(10):3962–71. pmid:34328684
  14. 14. Busque L, Buscarlet M, Mollica L, Levine RL. Concise review: age-related clonal hematopoiesis: stem cells tempting the devil. Stem Cells. 2018;36(9):1287–94. pmid:29883022
  15. 15. Kar SP, Quiros PM, Gu M, Jiang T, Mitchell J, Langdon R, et al. Genome-wide analyses of 200,453 individuals yield new insights into the causes and consequences of clonal hematopoiesis. Nat Genet. 2022;54(8):1155–66. pmid:35835912
  16. 16. Li X, Zheng Y, Zhu H, Lin X, Zhang Y, Zhao Y, et al. Risk of onset of hematological malignancies in patients infected with the hepatitis B virus: results from a large-scale retrospective cohort study in China. Acta Haematol. 2017;137(4):209–13. pmid:28514772
  17. 17. Ruiz-Magaña MJ, Martínez-Aguilar R, Lucendo E, Campillo-Davo D, Schulze-Osthoff K, Ruiz-Ruiz C. The antihypertensive drug hydralazine activates the intrinsic pathway of apoptosis and causes DNA damage in leukemic T cells. Oncotarget. 2016;7(16):21875–86. pmid:26942461
  18. 18. Dueñas-Gonzalez A, Coronel J, Cetina L, González-Fierro A, Chavez-Blanco A, Taja-Chayeb L. Hydralazine-valproate: a repositioned drug combination for the epigenetic therapy of cancer. Expert Opin Drug Metab Toxicol. 2014;10(10):1433–44. pmid:25154405
  19. 19. Candelaria M, Herrera A, Labardini J, González-Fierro A, Trejo-Becerril C, Taja-Chayeb L, et al. Hydralazine and magnesium valproate as epigenetic treatment for myelodysplastic syndrome. Preliminary results of a phase-II trial. Ann Hematol. 2011;90(4):379–87. pmid:20922525
  20. 20. Deng C, Lu Q, Zhang Z, Rao T, Attwood J, Yung R, et al. Hydralazine may induce autoimmunity by inhibiting extracellular signal-regulated kinase pathway signaling. Arthritis Rheum. 2003;48(3):746–56. pmid:12632429
  21. 21. Singh V, Sharma P, Capalash N. DNA methyltransferase-1 inhibitors as epigenetic therapy for cancer. Curr Cancer Drug Targets. 2013;13(4):379–99. pmid:23517596
  22. 22. Andrade F de O, Nguyen NM, Warri A, Hilakivi-Clarke L. Reversal of increased mammary tumorigenesis by valproic acid and hydralazine in offspring of dams fed high fat diet during pregnancy. Sci Rep. 2019;9(1):20271. pmid:31889127
  23. 23. Venugopal K, Feng Y, Shabashvili D, Guryanova OA. Alterations to DNMT3A in hematologic malignancies. Cancer Res. 2021;81(2):254–63. pmid:33087320
  24. 24. Stengel A, Kern W, Haferlach T, Meggendorfer M, Haferlach C. The 5q deletion size in myeloid malignancies is correlated to additional chromosomal aberrations and to TP53 mutations. Genes Chromosomes Cancer. 2016;55(10):777–85. pmid:27218649
  25. 25. Sasaki K, Kanagal-Shamanna R, Montalban-Bravo G, Assi R, Jabbour E, Ravandi F, et al. Impact of the variant allele frequency of ASXL1, DNMT3A, JAK2, TET2, TP53, and NPM1 on the outcomes of patients with newly diagnosed acute myeloid leukemia. Cancer. 2020;126(4):765–74. pmid:31742675
  26. 26. Desai P, Mencia-Trinchant N, Savenkov O, Simon MS, Cheang G, Lee S, et al. Somatic mutations precede acute myeloid leukemia years before diagnosis. Nat Med. 2018;24(7):1015–23. pmid:29988143
  27. 27. Abelson S, Collord G, Ng SWK, Weissbrod O, Mendelson Cohen N, Niemeyer E, et al. Prediction of acute myeloid leukaemia risk in healthy individuals. Nature. 2018;559(7714):400–4. pmid:29988082
  28. 28. Couronné L, Bastard C, Bernard OA. TET2 and DNMT3A mutations in human T-cell lymphoma. N Engl J Med. 2012;366(1):95–6. pmid:22216861
  29. 29. Sohn TA, Bansal R, Su GH, Murphy KM, Kern SE. High-throughput measurement of the Tp53 response to anticancer drugs and random compounds using a stably integrated Tp53-responsive luciferase reporter. Carcinogenesis. 2002;23(6):949–57. pmid:12082016
  30. 30. Errichiello E, Mina T, Morbini P, Zecca M, Zuffardi O. FANCA, TP53, and del(5q)/RPS14 alterations in a patient with T-cell non-Hodgkin lymphoma and concomitant Fanconi anemia and Li-Fraumeni syndrome. Cancer Genet. 2021;256–257:179–83. pmid:33183999
  31. 31. Lin L-Y, Warren-Gash C, Smeeth L, Chen P-C. Data resource profile: the National Health Insurance Research Database (NHIRD). Epidemiol Health. 2018;40:e2018062. pmid:30727703
  32. 32. Lee P-C, Kao F-Y, Liang F-W, Lee Y-C, Li S-T, Lu T-H. Existing data sources in clinical epidemiology: the Taiwan National Health Insurance Laboratory databases. Clin Epidemiol. 2021;13:175–81. pmid:33688263
  33. 33. Hsieh C-Y, Su C-C, Shao S-C, Sung S-F, Lin S-J, Kao Yang Y-H, et al. Taiwan’s National Health Insurance Research Database: past and future. Clin Epidemiol. 2019;11:349–58. pmid:31118821
  34. 34. Sung S-F, Hsieh C-Y, Lin H-J, Chen Y-W, Yang Y-HK, Li C-Y. Validation of algorithms to identify stroke risk factors in patients with acute ischemic stroke, transient ischemic attack, or intracerebral hemorrhage in an administrative claims database. Int J Cardiol. 2016;215:277–82. pmid:27128546
  35. 35. Kao W-H, Hong J-H, See L-C, Yu H-P, Hsu J-T, Chou I-J, et al. Validity of cancer diagnosis in the National Health Insurance database compared with the linked National Cancer Registry in Taiwan. Pharmacoepidemiol Drug Saf. 2018;27(10):1060–6. pmid:28815803
  36. 36. Unger T, Borghi C, Charchar F, Khan NA, Poulter NR, Prabhakaran D, et al. 2020 International society of hypertension global hypertension practice guidelines. Hypertension. 2020;75(6):1334–57. pmid:32370572
  37. 37. Verdecchia P, Reboldi G, Angeli F. The 2020 International Society of Hypertension global hypertension practice guidelines—key messages and clinical considerations. Eur J Intern Med. 2020;82:1–6. pmid:32972800
  38. 38. Unger T, Borghi C, Charchar F, Khan NA, Poulter NR, Prabhakaran D, et al. 2020 International Society of Hypertension global hypertension practice guidelines. J Hypertens. 2020;38(6):982–1004. pmid:32371787
  39. 39. Zolotareva O, Saik OV, Königs C, Bragina EY, Goncharova IA, Freidin MB, et al. Comorbidity of asthma and hypertension may be mediated by shared genetic dysregulation and drug side effects. Sci Rep. 2019;9(1):16302. pmid:31705029
  40. 40. Kale VP, Habib H, Chitren R, Patel M, Pramanik KC, Jonnalagadda SC, et al. Old drugs, new uses: drug repurposing in hematological malignancies. Semin Cancer Biol. 2021;68:242–8. pmid:32151704
  41. 41. Hess KR. Graphical methods for assessing violations of the proportional hazards assumption in Cox regression. Stat Med. 1995;14(15):1707–23. pmid:7481205
  42. 42. Austin PC, Fine JP. Practical recommendations for reporting Fine-Gray model analyses for competing risk data. Stat Med. 2017;36(27):4391–400. pmid:28913837
  43. 43. He L, Shi L, Du Z, Huang H, Gong R, Ma L, et al. Mebendazole exhibits potent anti-leukemia activity on acute myeloid leukemia. Exp Cell Res. 2018;369(1):61–8. pmid:29750898
  44. 44. Timmers GJ, Kessels LW, Wilhelm AJ, Veldkamp AI, Bosch TM, Beijnen JH, et al. Effects of cyclosporine a on single-dose pharmacokinetics of intravenous itraconazole in patients with hematologic malignancies. Ther Drug Monit. 2008;30(3):301–5. pmid:18520601
  45. 45. Song Y, Chen S, Xiang W, Xiao M, Xiao H. The mechanism of treatment of multiple myeloma with metformin by way of metabolism. Arch Med Sci. 2020;17(4):1056–63. pmid:34336033
  46. 46. Rücker FG, Lang KM, Fütterer M, Komarica V, Schmid M, Döhner H, et al. Molecular dissection of valproic acid effects in acute myeloid leukemia identifies predictive networks. Epigenetics. 2016;11(7):517–25. pmid:27309669
  47. 47. Xu M, Hou Y, Sheng L, Peng J. Therapeutic effects of thalidomide in hematologic disorders: a review. Front Med. 2013;7(3):290–300. pmid:23856973
  48. 48. Global Burden of Disease Cancer Collaboration, Fitzmaurice C, Abate D, Abbasi N, Abbastabar H, Abd-Allah F, et al. Global, regional, and national cancer incidence, mortality, years of life lost, years lived with disability, and disability-adjusted life-years for 29 cancer groups, 1990 to 2017: a systematic analysis for the Global Burden of Disease Study. JAMA Oncol. 2019;5(12):1749–68. pmid:31560378
  49. 49. Mills KT, Stefanescu A, He J. The global epidemiology of hypertension. Nat Rev Nephrol. 2020;16(4):223–37. pmid:32024986
  50. 50. Buscarlet M, Provost S, Zada YF, Barhdadi A, Bourgoin V, Lépine G, et al. DNMT3A and TET2 dominate clonal hematopoiesis and demonstrate benign phenotypes and different genetic predispositions. Blood. 2017;130(6):753–62. pmid:28655780
  51. 51. Prokocimer M, Unger R, Rennert HS, Rotter V, Rennert G. Pooled analysis of p53 mutations in hematological malignancies. Hum Mutat. 1998;12(1):4–18. pmid:9633814
  52. 52. Wang Y, Nguyen DT, Yang G, Anesi J, Kelly J, Chai Z, et al. A modified MTS proliferation assay for suspended cells to avoid the interference by hydralazine and β-mercaptoethanol. Assay Drug Dev Technol. 2021;19(3):184–90. pmid:33471568
  53. 53. Cervera E, Candelaria M, López-Navarro O, Labardini J, Gonzalez-Fierro A, Taja-Chayeb L, et al. Epigenetic therapy with hydralazine and magnesium valproate reverses imatinib resistance in patients with chronic myeloid leukemia. Clin Lymphoma Myeloma Leuk. 2012;12(3):207–12. pmid:22420986
  54. 54. Santos DMD, Da Silva EAP, Oliveira JYS, Marinho YY de M, Santana IR de, Heimfarth L, et al. The therapeutic value of hydralazine in reducing inflammatory response, oxidative stress, and mortality in animal sepsis: involvement of the PI3K/AKT pathway. Shock. 2021;56(5):782–92. pmid:33555842
  55. 55. Li C, Su Z, Ge L, Chen Y, Chen X, Li Y. Cardioprotection of hydralazine against myocardial ischemia/reperfusion injury in rats. Eur J Pharmacol. 2020;869:172850. pmid:31830459
  56. 56. Wu Q, Sun Z, Chen Z, Liu J, Ding H, Luo C, et al. The discovery of a non-competitive GOT1 inhibitor, hydralazine hydrochloride, via a coupling reaction-based high-throughput screening assay. Bioorg Med Chem Lett. 2022;73:128883. pmid:35820623
  57. 57. Cheng Z, Dai Y, Zeng T, Liu Y, Cui L, Qian T, et al. Upregulation of glutamic-oxaloacetic transaminase 1 predicts poor prognosis in acute myeloid leukemia. Front Oncol. 2020;10:379. pmid:32266153
  58. 58. Zhang Q, Lin Z, Yin X, Tang L, Luo H, Li H, et al. In vitro and in vivo study of hydralazine, a potential anti-angiogenic agent. Eur J Pharmacol. 2016;779:138–46. pmid:26968484
  59. 59. Lekovic D, Gotic M, Skoda R, Beleslin-Cokic B, Milic N, Mitrovic-Ajtic O, et al. Bone marrow microvessel density and plasma angiogenic factors in myeloproliferative neoplasms: clinicopathological and molecular correlations. Ann Hematol. 2017;96(3):393–404. pmid:27924369
  60. 60. Alexandrakis MG, Passam FH, Boula A, Christophoridou A, Aloizos G, Roussou P, et al. Relationship between circulating serum soluble interleukin-6 receptor and the angiogenic cytokines basic fibroblast growth factor and vascular endothelial growth factor in multiple myeloma. Ann Hematol. 2003;82(1):19–23. pmid:12574959
  61. 61. Feher A, Kampaktsis PN, Parameswaran R, Stein EM, Steingart R, Gupta D. Aspirin is associated with improved survival in severely thrombocytopenic cancer patients with acute myocardial infarction. Oncologist. 2017;22(2):213–21. pmid:28159866
  62. 62. Lu Y, Liu X-F, Liu T-R, Fan R-F, Xu Y-C, Zhang X-Z, et al. Celecoxib exerts antitumor effects in HL-60 acute leukemia cells and inhibits autophagy by affecting lysosome function. Biomed Pharmacother. 2016;84:1551–7. pmid:27884749
  63. 63. Lindner S, Krönke J. The molecular mechanism of thalidomide analogs in hematologic malignancies. J Mol Med (Berl). 2016;94(12):1327–34. pmid:27492707
  64. 64. Walter RB, Milano F, Brasky TM, White E. Long-term use of acetaminophen, aspirin, and other nonsteroidal anti-inflammatory drugs and risk of hematologic malignancies: results from the prospective Vitamins and Lifestyle (VITAL) study. J Clin Oncol. 2011;29(17):2424–31. pmid:21555699
  65. 65. Cerhan JR, Anderson KE, Janney CA, Vachon CM, Witzig TE, Habermann TM. Association of aspirin and other non-steroidal anti-inflammatory drug use with incidence of non-Hodgkin lymphoma. Int J Cancer. 2003;106(5):784–8. pmid:12866040
  66. 66. White WB, Kent J, Taylor A, Verburg KM, Lefkowith JB, Whelton A. Effects of celecoxib on ambulatory blood pressure in hypertensive patients on ACE inhibitors. Hypertension. 2002;39(4):929–34. pmid:11967252
  67. 67. Bautista LE, Vera LM. Antihypertensive effects of aspirin: what is the evidence?. Curr Hypertens Rep. 2010;12(4):282–9. pmid:20524092
  68. 68. Wang L, Zhang R, Yu L, Xiao J, Zhou X, Li X, et al. Aspirin use and common cancer risk: a meta-analysis of cohort studies and randomized controlled trials. Front Oncol. 2021;11:690219. pmid:34277434
  69. 69. McNeil JJ, Gibbs P, Orchard SG, Lockery JE, Bernstein WB, Cao Y, et al. Effect of aspirin on cancer incidence and mortality in older adults. J Natl Cancer Inst. 2021;113(3):258–65. pmid:32778876
  70. 70. Okabe S, Tauchi T, Tanaka Y, Sakuta J, Ohyashiki K. Anti-leukemic activity of axitinib against cells harboring the BCR-ABL T315I point mutation. J Hematol Oncol. 2015;8:97. pmid:26239229
  71. 71. Kadowaki H, Ishida J, Akazawa H, Yagi H, Saga-Kamo A, Umei M, et al. Axitinib induces and aggravates hypertension regardless of prior treatment with tyrosine kinase inhibitors. Circ Rep. 2021;3(4):234–40. pmid:33842729
  72. 72. Copland E, Canoy D, Nazarzadeh M, Bidel Z, Ramakrishnan R, Woodward M, et al. Antihypertensive treatment and risk of cancer: an individual participant data meta-analysis. Lancet Oncol. 2021;22(4):558–70. pmid:33794209
  73. 73. Cameron HA, Ramsay LE. The lupus syndrome induced by hydralazine: a common complication with low dose treatment. Br Med J (Clin Res Ed). 1984;289(6442):410–2. pmid:6432120
  74. 74. Collins KS, Raviele ALJ, Elchynski AL, Woodcock AM, Zhao Y, Cooper-DeHoff RM, et al. Genotype-guided hydralazine therapy. Am J Nephrol. 2020;51(10):764–76. pmid:32927458
  75. 75. Sun X, Yang S, Song W. Prazosin inhibits the proliferation and survival of acute myeloid leukaemia cells through down-regulating TNS1. Biomed Pharmacother. 2020;124:109731. pmid:31954876
  76. 76. Espinoza-Zamora JR, Labardini-Méndez J, Sosa-Espinoza A, López-González C, Vieyra-García M, Candelaria M, et al. Efficacy of hydralazine and valproate in cutaneous T-cell lymphoma, a phase II study. Expert Opin Investig Drugs. 2017;26(4):481–7. pmid:28277033