Risk of cancer history in cardiovascular disease among individuals with hypertension | Hypertension Research – Nature.com

Study population

This retrospective observational cohort study used the JMDC Claims Database (JMDC Inc., Tokyo, Japan) from January 2005 through May 2022 [16,17,18]. The JMDC Claims Database is a nationwide database that incorporates information on annual health checkups (such as blood pressure [BP] measurement and anthropometric measurements) and individual health insurance records from more than 60 insurers in Japan. The majority of registered individuals were employees who work for relatively large companies in Japan under the coverage of kempo, a health insurance system for employees (i.e., working-age population). Because the database includes administrative claims data from both outpatient and inpatient settings, researchers can follow individual patients even if they move to a different hospital during medical treatment. Diagnosis data (confirmed or suspected) in the database are curated according to the International Classification of Diseases, 10th Revision (ICD-10).

For the present study included 999,635 individuals aged 1874 years with hypertension, more than 1 year after insurance enrollment. We defined hypertension was defined as systolic blood pressure (SBP)140mmHg and diastolic blood pressure 90mmHg or requiring the use of the following antihypertensive medicationsWorld Health Organization Anatomical Therapeutic Chemical (WHO-ATC) codes: C02, C03, C04, C07, C08, and C09. Study participants were censored if they reached the age of 75 years or older. Individuals were excluded if they had a history of CVD, including myocardial infarction (MI), angina pectoris (AP), stroke, heart failure (HF), or atrial fibrillation (AF) (n=136,576), history of renal replacement therapy (n=868), or there were missing data regarding cigarette smoking (n=33,157), alcohol consumption (n=54,630), and physical inactivity (n=26,784). Finally, a total of 747,620 patients were included in this study (Supplementary Fig.1).

This study was approved by the Ethics Committee of the University of Tokyo (approval number: 2018-10862) and conducted in accordance with the principles of the Declaration of Helsinki. Because the JMDC Claims Database included anonymized data, the requirement for informed consent was waived.

A history of cancer was defined as being confirmed diagnosis with malignant neoplasms (ICD-10 codes: C00-D09) before the index date (i.e., initial health check-up). Additionally, we extracted information regarding the diagnosis of certain cancer types which are reportedly more prevalent in Japanese adults (https://ganjoho.jp/reg_stat/statistics/stat/summary.html). Diagnosis codes of each cancer based on the ICD-10 code are shown in Supplementary Table1.

Obesity was defined as a body mass index of 25kg/m2. Diabetes mellitus was defined as a fasting glucose level of 126mg/dL or the use of blood glucose-lowering medications (including insulin). Dyslipidemia was defined using the following criteria: (1) low-density lipoprotein cholesterol level 140mg/dL; (2) high-density lipoprotein cholesterol level <40mg/dL; (3) triglyceride level 150mg/dL; or (4) the use of lipid-lowering medications. Information regarding smoking status (current or previous/never) and alcohol consumption (every day or not every day) was obtained by a self-reported questionnaire filled out during the health checkup. Lastly, self-reported physical inactivity was defined as not performing a minimum of 30min of exercise at least two times a week or not walking for more than 1h per day.

The primary outcome was the incidence of composite CVD events, including MI (ICD-10 codes: I21.0I21.4 and I21.9), AP (ICD-10 codes: I20.0, I20.1, I20.8, and I20.9), stroke (ICD-10 codes: I63.0, I63.1I63.6, I63.8, I63.9, I60.0I61.1, I61.3I61.6, I61.9, I62.9, and G45.9), HF (ICD-10 codes: I50.0, I50.1, I50.9, and I11.0), and AF (ICD-10 codes: I48.0, I48.1, I48.2, I48.3, I48.4, and I48.9) [17, 19]. We also evaluated the individual incidence of MI, AP, stroke, HF, and AF as secondary outcomes. We used confirmed diagnosis records to define each incident of CVD, thereby excluding any suspected diagnoses from the analysis. Study participants were followed until the incidence of outcomes, dropped from insurance coverage, death, or study end date (May 2022).

We used median (2575th percentiles) or number (percentage) to describe the basic characteristics of study participants stratified by cancer history. For comparing basic characteristics between the two groups, we used the Chi-square test for categorical variables and the Mann-Whitney U test for continuous variables. Additionally, we used the KaplanMeier curves with a log-rank test to estimate the incidence of each CVD event according to cancer history. We also constructed a Cox proportional hazard regression model to examine the association between a history of cancer and CVD incidence. Model 1 was an unadjusted model, so we adjusted for age and sex in Model 2; Model 3 was further adjusted for SBP, obesity, diabetes mellitus, dyslipidemia, cigarette smoking, alcohol consumption, and physical inactivity. We also computed the hazard ratios (HRs) for each covariate in the multivariable Cox proportional hazard regression (Model 3).

Furthermore, individuals were categorized into three groups: those without a history of cancer, those with a history of cancer but did not receive active chemotherapy, and those with a history of cancer who received active chemotherapy. Active chemotherapy was defined as the use of antineoplastic agents (WHO-ATC codes: L01) within 3 months before the index date because the Japanese system of universal health insurance allows for a maximum prescription term of 3 months. Using this categorization, we constructed a Cox proportional hazard regression to examine the association between a combination of cancer history and chemotherapy with CVD incidence. We assessed the HRs using Cox proportional hazard regression for composite CVD events among individuals with the five most prevalent cancer types in our dataset.

Additionally, we conducted seven sensitivity analyses. First, we set an induction period of 1 year to exclude individuals with latent CVD. Second, we constructed the Fine-Gray competing risk regression model by accounting for death as a competing risk. Third, we imputed missing data on covariates using multiple imputations by chained equations and 20 iterations, assuming to be missing at random for covariates. Fourth, we performed a subgroup analysis stratified by sex (men, women), age (50 years, <50 years), SBP (median SBP [140mmHg], <140mmHg), and the use of antihypertensive medications (with, without). Fifth, we created a separate cohort after 1:1 matching individuals with a history of cancer and those without a history of cancer according to age (every 5 years), sex, and smoking. Sixth, we excluded cigarette smoking from covariates. Seventh, we repeated the primary analyses among individuals aged 60 years. A two-sided p value of <0.05 was considered statistically significant. All analyses were performed using the Stata software (version 17; StataCorp LLC, College Station, TX, USA).

Read this article:
Risk of cancer history in cardiovascular disease among individuals with hypertension | Hypertension Research - Nature.com