Carvedilol use is associated with reduced cancer risk: A nationwide population-based cohort study.

Carvedilol use is associated with reduced cancer risk: A

nationwide

population-based cohort study

Chin-Sheng Lin

,Wei-Shiang Lin

, Cheng-Li Lin

, Chia-Hung Kao

⁎

1. Introduction

The β-adrenergic signaling pathway mediates sympathetic nerve system-induced fght-or-fight stress responses, which is suggested to regulate the initiation and progression of cancer through several

mechanisms, such as infammation, angiogenesis, resistance to apoptosis, stimulation of tumor cell invasion, and epithelial–mesenchymal

transition [1]. Previous studies have demonstrated that the use of propranolol,

a nonselective β-blocker, exerts antitumor effects through

antiangiogenesis, inhibition of stress-induced macrophage infltration, and hypoxia-inducible factor 1α-mediated inhibition of vascular endothelial

growth factor (VEGF) signaling [2–5].Moreover,when combined

with gemcitabine, a β-blocker is known to exert synergistic antiproliferative and apoptotic effects through inhibition of nuclear factor (NF)-κB

signaling [6]. These studies indicated the possible role of β-adrenergic

signaling in tumor biology.

Carvedilol, a nonselective β-blocker with α1 adrenoceptor blocking and potent antioxidant activity, is currently being used for the treatment of coronary artery disease (CAD), hypertension, or symptomatic

heart failure [7,8]. Data fromseveral in vitro studies have demonstrated

that carvedilol exerts antiproliferative and cytotoxic effects on various

human cancer cell lines [9–11]. A combination of carvedilol and imatinib

mesylate revealed synergistic tumor cytotoxicity in rat glioma cells

[9]. In addition, compared with other β-blockers, carvedilol exhibited

the most potent antiproliferative and antiangiogenic properties in

neuroblastoma cells [2]. Carvedilol enhanced response to chemotherapy

with vincristine, which provided optimal survival benefts among neuroblastoma-bearing mice treated with vincristine and

other β-blockers [2].

Although there is strong preclinical evidence suggesting the benefcial role of carvedilol in cancer treatment, no large population-based study has evaluated the association between carvedilol and cancer

of cancer in a large population-based cohort study.We used the Taiwan National Health Insurance Database for analysis and statistics,

and hypothesized that treatment with carvedilol reduces the risk of cancer. We suggested that the long-term use of carvedilol is associated with reduced upper gastrointestinal tract and lung cancer risk, indicating that carvedilol could be a potential therapeutic agent in these cancers' prevention.

2. Methods 2.1. Data source

The National Health Insurance (NHI) program was initiated by the Taiwanese government in 1995 to provide comprehensive health care for all residents of Taiwan. More than 99% of the 22.6 million residents

of Taiwan are enrolled in this system (http://www.nhi.gov.tw/english/

index.aspx). The Bureau of National Health Insurance releases scrambled data to the National Health Research Institutes (NHRI) to establish the National Health Insurance Research Database (NHIRD). This study adopted the Longitudinal Health Insurance Database 2000 (LHID2000), a subset of one million insurants randomly selected from the NHIRD from 1996 to 2000. The LHID2000 contains medical data identifed by encrypted patient numbers, including sex, date of birth, registry ofmedical services, diagnoses (identifed based on International Classifcation of Diseases, Ninth Revision, Clinical Modifcation [ICD-9-CM] diagnostic and procedure codes), and medication prescription details. This study was approved by the Institutional Review Board of China Medical University (CMU-REC-101-012).

2.2. Sampled patients

This retrospective cohort study used data extracted from the LHID2000. The current study included patients aged N20 years for whom complete information was available with regard to age and sex and who had no history of cancer (ICD-9-CM codes 140–208) between January 1, 2000 and December 31, 2010. The patients were divided into 2 cohorts based on their carvedilol use: the carvedilol cohort included patients who had received carvedilol therapy for at least 2 years (730 d); the noncarvedilol cohort included patients who had not received carvedilol therapy before and during follow-up. The index date in the carvedilol cohort was the 730th day, and the index date in the noncarvedilol cohort (controls) was the same date as in the carvedilol cohort. Patients in the carvedilol and noncarvedilol cohorts were

selected by 1:1 matching based on a propensity score [12]. The propensity score was calculated using a logistic regression to estimate the

probability of the treatment assignment, based on the baseline variables including year of receiving carvedilol treatment, age, sex, Charlson comorbidity index (CCI) score and comorbidities of hypertension (ICD-9-CM codes 401–405), CAD (ICD-9-CM codes 410–414), atrial fbrillation (AF; ICD-9-CM code 427.3), supraventricular tachycardia (ICD-9-CM codes 427.0, 427.2), and palpitation (ICD-9-CM code

785.1). The C-statistic of the logistic regression modelwas 0.51. The carvedilol cohort comprised 6771 patients and the noncarvedilol cohort

comprised 6771 patients. The cumulative censoring rate over 12 years (2000–2011) was 13.54% in the carvediolol cohort, which was slightly lower than that in the noncarvediolol cohort (17.5%). The possible reasons for the discontinuity of national health insurance include death, withdrawal of insurance, immigration, prison sentence, etc. 2.3. Outcome measurement

The study outcome was a diagnosis of cancer (ICD-9-CM codes 140–195, 200–208) during the 11-year follow-up. Cancer events were identifed according to the Registry of Catastrophic Illness Patient

Database,which is a subset of the NHIRD. Cancer is categorized as a catastrophic illness in the NHI programand patients newly diagnosed with

cancer can apply for a catastrophic illness certifcation. The certifcation is issued by the government after a stringent process of verifcation, which involves review of medical records, images, and pathology reports by a panel of specialists and experts on the disease. The study patients were followed up from the index date until the date of cancer diagnosis, withdrawal from the NHI program, or when the database ended (December 31, 2011).

2.4. Statistical analysis

The distribution of demographic status and comorbidities was expressed as a frequency (percentage) or mean ± standard deviation (SD). The categorical variables were analyzed using the chi-square test, and the continuous variables of the baseline characteristics of the carvedilol and noncarvedilol cohorts were analyzed using the Student t test. The Kaplan–Meiermethodwas employed to determine the cumulative incidence curves of cancer between the 2 cohorts, and the logrank

test was conducted to examine the difference between the curves. The incidence densities were calculated for each cohort by sex, age, and

type of cancer. Cox proportional hazards model stratifcation of the matched pairs was used to measure the effect of carvedilol on the time to cancer diagnosis. The hazard ratio (HR) is presented with 95% or 99.5% confdence interval (CI). The Bonferroni adjustment was used in multiple comparisons. All statistical analyses were performed using SAS, Version 9.3 (SAS Institute, Cary, NC, USA). Comparison results with a P value less than .05 were considered statistically signifcant. 3. Results

Eligible study patients included 6771 patients who used carvedilol for 2 years (carvedilol cohort) and 6771 patients who had never used carvedilol therapy, matched on the propensity score (noncarvedilol

cohort or controls) (Table 1). Themean age of the patients in the carvedilol

and noncarvedilol cohorts was 62.8 (±12.8) and 63.1 (±12.9)

years, respectively. No signifcant differences in distribution of age, sex, CCI score, and comorbidities between the carvedilol and noncarvedilol cohorts were found. The mean follow-up period was 5.17 (SD = 2.78) and 4.93 (SD=2.91) years in the carvedilol and noncarvedilol cohorts, respectively (data not shown). The plot of the Kaplan–Meier analysis

showed that, by the end of the 11-year follow-up period, the cumulative incidence of cancer was lower for the carvedilol cohort than for the

noncarvedilol cohort (log-rank test: p b .001) (Fig. 1).

The overall sex- and age-specifc incidence rates and HR of these two

cohorts are shown in Table 2. The overall incidence density of cancer

was signifcantly lower in the carvedilol cohort than in the noncarvedilol cohort (9.77 vs 11.9 per 1000 person-y). The patients in the

carvedilol cohort had a 26% lower cancer risk compared with patients in the noncarvedilol cohort (95% CI=0.63–0.87). The incidence density of cancer was higher in men than in women for both cohorts. The sexspecifc carvedilol to noncarvedilol relative risk was lower for both

women (HR = 0.73; 95% CI = 0.56–0.94) and men (HR = 0.75; 95% CI=0.61–0.92). The incidence of cancer increased with age in both cohorts, and the age-specifc carvedilol to noncarvedilol relative risk was

lower for all age groups. Table 3 shows the analyses of specifc cancer

types in the carvedilol and noncarvedilol cohorts. We observed that subjects using carvedilol showed a decreased risk of upper gastrointestinal (GI) tract cancer (HR = 0.30, 99.5% CI = 0.10–0.87). The risk of

lung cancerwas also signifcantly lower for patients in the carvedilol cohort than in the noncarvedilol cohort (HR=0.59, 95% CI=0.37–0.94).

In addition, the duration of carvedilol use was associated with the

risk of upper GI tract cancer, and lung cancer. Table 4 shows the risk

of upper GI tract cancer and lung cancer stratifed according to the duration of carvedilol use. The risk of upper GI tract cancer decreased when

the patients used carvedilol for longer than 1095 days. 4. Discussion

In this nationwide population-based retrospective cohort study, we observed a signifcant 26% reduction of cancer risk in patients receiving long-term carvedilol treatment compared with the noncarvedilol cohort, irrespective of age, sex, and baseline comorbidities such as hypertension, coronary artery disease, and AF. When we focused on certain cancer types, patients treated with carvedilol had a signifcant reduction of upper GI tract (HR = 0.30; 99.55% CI = 0.10–0.87) and lung (HR = 0.59; 95% CI = 0.37–0.94) cancer risks, compared with patients in the noncarvedilol cohort. These results indicated the benefcial effects of carvedilol in cancer prevention, particularly upper GI tract and lung cancers.

The NHIRD used in the present investigation is an effective database to provide population-based studieswith age- and sex-matched groups. Because participation in the NHI program is mandatory and all

Taiwanese residents can access medical care with low copayments, the loss to follow-up is low. Moreover, prescription of carvedilol is based on complete clinical electrocardiographic and echocardiographic assessments. Pathologic proof of malignancy is necessary to apply for a cancer catastrophic illness certifcate to obtain exemption from related medical costs such as hospital expenses; therefore, data regarding the use of carvedilol and cancer diagnosiswere reliable in the current study. As we mentioned above, carvedilol as well as other types of

β-blockers are being used to treat chronic stable disease, such as CAD,

hypertension, and heart failure [7,8]. This suggests that long-term use

of carvedilol is not a big concern regarding its safety issue. Moreover, although the most common adverse reactions of carvedilol were edema, dizziness, bradycardia, and hypotension, most patients tolerate it well even in chronic use. Similar to other types of β-blockers, only

5% of patients who taking carvedilol fnally discontinued it [13]. Such

evidence elucidates the safety and effcacy of carvedilol in its longterm use.

cancer risk, previous studies have suggested that carvedilol exhibits anticancer activities by arresting cells at the G0/G1 phase, inducing apoptosis and causing severe mitochondria damage in glioma cells

[9]. In addition to antiproliferative and antiangiogenic effects, carvedilol

reverses multidrug resistance to anticancer drugs through Pglycoprotein inhibition, which can enhance response to chemotherapy

[2,9,14]. Moreover, compared with other β-blockers, carvedilol

exerts potent and multifaceted therapeutic effects including antioxidant and anti-infammatory properties. Carvedilol, but not propranolol, inhibits intracellular reactive oxygen species production, tumornecrosis-factor-alpha-stimulated NF-κB and activator protein-1 transcription factors activation, and infammatory cytokine expression in

human mononuclear cells [15–18]. Additional in vivo investigations

have suggested that carvedilol reduces matrix metalloproteinase

(MMP)-2 and MMP-9 expression, likely through modulation of redoxrelated

pathways [19]. This evidence supports and elucidates the anticancer

effects of carvedilol.

Observational studies have revealed controversial clinical results

of β-blocker use in cancer patients. Inhibition of the β2-adrenergic pathway

has been shown to reduce breast cancer progression and mortality

[20]. Furthermore, treatment with β-blockers provides clinical benefts

to patients with breast, melanoma, and ovarian cancers, or non-small

cell lung cancer (NSCLC) [21–24]. Unlike the benefcial roles of

β-blockers in cancer patients, a large cohort study suggested that the use of β-blockers is not associated with a decreased risk of prostate

cancer and all-cause mortality [25]. McCourt et al. suggested that the

use of β-blockers after the diagnosis of malignantmelanomais not associatedwith reduced risk of death frommelanoma in a population-based

study conducted in the United Kingdom [26]. Given these potentially disputable fndings, the role of

β-blockers in cancer prevention remains

unclear. These controversial results might be due to various cancer types and antitumor activities of different β-blockers.

We demonstrated that the long-term use of carvedilol is associated

with reduced upper GI tract cancer risk (Table 3). Although the relative

HRon upper GI tract cancer is higher in patientswho took carvedilol less than 1095 days than those who did not take carvedilol, this trend did not reach statistical signifcance. Interestingly, there are signifcantly benefcial effects of carvedilol on upper GI tract cancer risk in patients

treated with carvedilol for longer than 1,095 days, indicating the dosedependent

effects of carvedilol on upper GI tract cancer (Table 4). Previous

report suggested that the β-adrenergic signaling pathway mediates chloroform extract of cigarette smoke induced proliferation of human

esophageal squamous cell carcinoma cells [27]. Additionally, other

β-blocker, such as propranolol has been reported to induce apoptosis and repress stomach cancer cell growth through the inhibition of the downstream NF-κB, VEGF, MMPs, and cyclooxygenase-2 pathways

[28]. These results provide strong evidence of the therapeutic potential

of carvedilol on upper GI tract cancers.

In addition to upper GI tract cancer, the risk of lung cancerwas lower

in patients treatedwith carvedilol (Table 3). Although the effects of carvedilol

on lung cancer have not been reported, previous in vitro studies

have suggested that exposure to nicotine through tobacco smoke or nicotine supplements facilitates the growth and progression of NSCLC,

which might be suppressed by propranolol [29]. Furthermore, Wang

et al. indicated that the use of β-blockers improved survival outcome

in NSCLC patients receiving radiation therapy [24]. The results indicate

that β-blockers may reduce the risk of NSCLC among smokers and could be used to improve the clinical outcome of standard cancer therapy. However, since reports related to carvedilol and lung cancer are

scant, and dose effect of carvedilol use on lung cancer could not be observed

in the current study (Table 4), the effects of carvedilol on lung

cancer must be confrmed.

The present study did not reveal a signifcant difference in reduction

of colon cancer risk in patients who received carvedilol treatment comparedwith

thosewho did not, as shown in Table 3. Jansen et al. demonstrated

that β-blocker use is not associated with a decreased risk of

colorectal cancer [30], which is in agreement with our fndings. In the

Jansen et al. case–control study, 1762 colorectal cancer patients and 1708 control patients fromGermanywere included,whowere followed up for 5 years. Despite analysis of subclasses of β-blockers, treatment with cardioselective or nonselective β-blockers, such as carvedilol, did not reduce the risk of colorectal cancer. These results suggest that carvedilol might not be helpful in reducing colorectal cancer risk.

Although a strength of this study was the use of a nationwide population-based cohort longitudinal analysis of the risk of cancer in Asian patients treated with carvedilol, several limitations should be

considered before interpreting the fndings. The NHIRD does not provide detailed information on the lifestyle or health-related factors of

patients, such as smoking, alcohol consumption, body mass index,

socioeconomic status, and family history of malignancy, which can increase the risk of cancer and were potential confounding factors in

this study. Moreover, evidence from a cohort study is generally considered to be of lower methodological quality than that from randomized

trials. An additional limitation is that we were unable to evaluate the

regular use of carvedilol. Only the information regarding the use of carvedilol or not is provided by the NHIRD. Therefore, the dose–response

relationship is demonstrated by cumulative duration of use, as indicated in Table 4. Although patients treated with carvedilol for b2 years were excluded from our analysis, lifetime exposure could not be estimated, and the cumulative duration of use may have been underestimated. Finally, despite the scrupulous study design to control the confounding factors, a key limitation of this study was its potential bias resulting from possible unmeasured or unknown confounders. However, as our previously published work regarding the effects of benzodiazepine on

cancer risk [31], the current study provides valuable information

regarding the effects of carvedilol on the reduction of cancer risk even with these limitations.

In conclusion, we demonstrated for the frst time that carvedilol is associated with reduced certain types of cancer and overall cancer risk in a large population-based cohort study. This study revealed the therapeutic potential of carvedilol in upper GI tract and lung cancers prevention.

Additional prospective randomized studies as well as in vivo and in vitro studies are necessary to verify the effects of carvedilol on cancer