The Family Heart Study: rationale, design and methodology

Cardiovascular disease (CVD) remains a global health burden and a leading cause of mortality and healthcare costs.¹ Aotearoa New Zealand, despite its small population, contributes disproportionately to global CVD-related deaths, Disability-Adjusted Life Years and expenditure.² Over the past seven decades, overall rates have declined due to fewer late-onset cases, but early-onset CVD—defined as onset before age 50 in men and 60 in women—is rising worldwide,³ a trend also seen in New Zealand with plateauing prevalences of early-onset CVD.² Up to 75% of premature cases are preventable with timely intervention.⁴

CVD risk begins accumulating early, potentially from age 20.⁵ Yet CVD risk assessments typically start at age 40 or later. In New Zealand, screening is recommended from age 45 for men and 55 for women, with earlier assessment in higher-risk groups, including Māori, Pacific peoples and those from the Indian subcontinent.⁶ Tools like PROCAM, Framingham and the New Zealand–specific PREDICT score incorporate clinical factors including family history.⁷ Recent studies suggest that cardiac biomarkers such as troponins, N-terminal pro–B-type natriuretic peptide (NT-proBNP) and creatine kinase MB could improve prediction.⁸ Genetic risk scores (GRS) may offer additional insights, particularly for those genetically predisposed.⁹ Only recently have studies begun to evaluate the ability of GRS to predict incident or recurrent CVD independent of traditional risk factors.¹⁰

Family history reflects shared environmental and genetic influences. Monogenic drivers of CVD, such as cardiomyopathies, inherited arrhythmias and familial hypercholesterolaemia (FH), are rare in the general population but common in affected families.¹¹ Monogenic FH, for example, increases CVD risk by up to 850%, often due to mutations in LDLR, APOB or PCSK9.¹² In contrast, polygenic risk arises from many common variants, each with a small effect, but collectively accounting for 30–50% of CVD risk.¹³ Many such variants are in non-coding regions and may influence traits like body mass index (BMI) or blood pressure via unknown mechanisms. However, evidence of many genetic influences has largely been derived from Europeans.¹⁴ Some individuals from affected families show no major risk factors or known monogenic variants yet have high sub-clinical disease burden.¹⁵ We hypothesise that these families carry i) a greater burden of common CVD-associated variants identified in recent genome-wide association studies, or ii) a few rare variants with a strong effect on risk, or iii) a combination of i and ii.

This paper describes the rationale, design and methodology behind the Family Heart Study (FHS), and recruitment of the first 28 participants. The FHS aims to investigate genetic variation associated with cardiovascular risk in New Zealanders with a personal and family history of premature CVD. The study utilises genetic, epigenetic, clinical, lifestyle and biomarker analyses to explore the complex relationship between family history, cardiac biomarkers, CVD risk factors and the risk of adverse clinical outcomes. The broad themes of the proposed research are to i) determine whether New Zealanders with a personal and family history of premature CVD have a greater burden of common or rare CVD-risk single-nucleotide polymorphisms (SNPs), and ii) determine whether genetic information has the potential to improve cardiovascular risk prediction in this group beyond established PREDICT CVD risk scores.

Methods

Study participants

The FHS, led by the Christchurch Heart Institute (CHI), commenced in 2010 at the University of Otago, Christchurch, New Zealand. Each recruited individual represents a separate family. The study was granted ethics approval from the Southern Health and Disability Ethics Committee (ref: 2024 AM 2948).

Participants were selected through a variety of screening strategies (Figure 1), including:

• screening emergency department admission lists;
• phoning patients from existing cohorts recruited by the CHI;
• screening the cardiac catheterisation lab admission list, Christchurch Hospital; and
• enlisting coronary disease patients from cardiology rehabilitation, Christchurch Hospital.

Inclusion criteria

Volunteers who met the inclusion criteria were contacted by phone and invited to attend a clinic appointment. An individual had to meet all of the following criteria:

1. European or Māori ethnicity (self-reported ethnicity of the participant and all four grandparents);
2. experienced a documented (confirmed via hospital or general practitioner medical records) early-onset coronary event (acute coronary syndrome or coronary revascularisation) at ≤50 years in men and ≤60 years in women;
3. had at least one first-degree relative who experienced a documented early-onset coronary event fitting the same criteria above; and
4. was able to provide signed informed consent.

Exclusion criteria

Individuals were excluded if no documentation was available confirming the age of onset and/or history of coronary events in a first-degree relative.

Verification of medical history

A comprehensive process was undertaken to verify the relative’s medical history. This involved direct contact with the relative, thorough examination of medical records and communication with general practitioners, including those overseas where necessary, to corroborate the reported age at the time of the first cardiovascular event.

Participant withdrawals

Participants could withdraw at any time. No participants have withdrawn to date.

Biological samples and research data for FHS

Measurements of blood pressure, weight, height and waist and hip circumferences were recorded, and body composition analysis was conducted using bioimpedance analysis. Additionally, participants completed a voluntary questionnaire regarding personal and first-degree family medical history and lifestyle factors (Figure 2). All data are entered into a secure, dedicated database. Participants were asked to provide a blood sample (approximately 70mL) for the extraction and analysis of deoxyribonucleic acid (DNA) and the measurement of heart hormones, and other markers. Blood was collected into chilled ethylenediaminetetraacetic acid tubes, stored on ice and centrifuged under standard conditions. Plasma was stored at −80 degrees Celsius at the CHI’s Health and Disability Ethics Committee–accredited tissue bank.

View Figure 1–2, Table 1–2.

Measurement of circulating biomarkers

Plasma levels of NT-proBNP,¹⁶ B-type natriuretic peptide (BNP),¹⁶ N-terminal pro–atrial natriuretic peptide (NT-proANP)¹⁷ and atrial natriuretic peptide (ANP)¹⁸ were measured in our laboratory using well-established immunoassays.

DNA extraction and genotyping

Genomic DNA was isolated from 3mL stored frozen whole blood or buffy coat samples using an automated KingFisher® Flex 24 instrument (Thermo Fisher Scientific, Waltham, United States of America [USA]) and DNA isolation kits (Macherey-Nagel, Düren, Germany). DNA samples were analysed for genetic variants, including but not limited to SNPs, copy number variants and other sequence variants, or altered DNA methylation, to assess their association with CVD risk.

Genotyping includes techniques ranging from analysis of individual SNPs via real-time quantitative polymerase chain reaction (PCR), to large-scale SNPs analysis using gene chip–based technologies, to whole exome or whole genome sequencing (including long-read nanopore sequencing). To date, genotyping of approximately 200,000 SNPs of relevance to CVD has been performed on DNA from 21 out of 28 FHS participants using Cardio-MetaboChip genotyping arrays (Illumina, San Diego, USA). The Cardio-MetaboChip is a cost-effective, custom, genotyping array with a high genotype call rate that is designed to detect SNPs linked to traits associated with cardiovascular risk.¹⁹ The arrays use Infinium chemistry and were run on an iScan instrument (Illumina) at AgResearch, Invermay, New Zealand.

DNA methylation analysis

Genomic DNA samples (n=28) were bisulphite-treated using EZ DNA methylation Gold Kits (Zymo Research, Irvine, USA), using specific incubation conditions for the Infinium Methylation Assay (Illumina). The Infinium HumanMethylation450 BeadChip array assesses methylation at over 450,000 cytosine-phosphate-guanine (CpG) sites using two assay types.²⁰ Completeness of bisulphite conversion was checked by PCR using specific primers with the Zymo Research Universal Methylation Human DNA Standard and Control Primers kit. Samples were run on HumanMethylation450 BeadChip arrays (Illumina) on an iScan instrument (Illumina) at AgResearch, Invermay, New Zealand.

Clinical outcomes

At recruitment, participants consented for the research team to access their medical records, obtaining prior medical history and medications from hospital records. To evaluate the prognostic power of genetic and other biomarkers under study, the research team will follow participants’ health for up to 20 years through linkage to administrative health data. Clinical outcomes including hospital admissions, mortality and cause of death data will be obtained from the administrative records of the Ministry of Health – Manatū Hauora and hospital medical datasets.

Clinical outcomes will be classified using International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification (ICD-10-AM) codes. In-hospital procedures will be classified with Australian Classification of Health Interventions codes. The primary outcome is major adverse cardiac events, defined as CVD events occurring after recruitment according to the New Zealand PREDICT definition.⁷ Secondary outcomes include fatal or non-fatal  acute coronary syndromes, fatal or non-fatal heart failure, fatal or non-fatal stroke, cardiovascular death and all-cause death. Cardiovascular death will be defined by ICD-10-AM codes or death within 28 days of cardiovascular hospitalisation.

Healthy control cohort

The existing Healthy Volunteers (HVOL) cohort²¹ is available as a control group for matching to the FHS. Participants (n=3,358, trial registry ACTRN12605000448640) were randomly selected from the Canterbury, New Zealand electoral rolls. Participants were aged 24–97 years at recruitment with no history of CVD. Participants completed a health and lifestyle questionnaire. Blood pressure, height, weight and waist and hip measurements were documented. Blood samples were taken at recruitment for neurohormone and genetic analyses. The study was approved by the Upper South A Ethics Committee (reference no. CTY/01/05/062), and all participants provided written, informed consent.

Within the HVOL cohort, 1,178 participants with 10-year follow-up free of cardiovascular events were genotyped using the Thermo Fisher Axiom Precision Medicine Diversity Array (PMDA; Thermo Fisher Scientific).

Coronary Disease Cohort Study (CDCS)

Another existing cohort, the CDCS,²² comprises participants (n=2,140) recruited from Christchurch and Auckland hospitals between 2002 and 2009 following admission with unstable angina or myocardial infarction. The cohort includes adults aged 19–94 years with established CVD. Clinical data and blood samples were collected at recruitment for biomarker and genetic analyses. The study was approved by the relevant ethics committee (reference no. CTY/02/02/018), and all participants provided written, informed consent.

A subset of participants (n=1,935) underwent genotyping for inclusion in genetic analyses using the PMDA. Among the genotyped CDCS participants, the cohort was stratified into three groups for comparison with the FHS: i) patients with premature CVD and self-reported family history of premature CVD (“FHS-like”, n=49), ii) patients with premature CVD without a family history of premature CVD (n=495), and iii) patients with CVD without a personal or family history of premature CVD (n=1,391).

Statistical approach

SNPs data (approximately 200,000 variants from the Cardio-MetaboChip) will undergo quality control in PLINK software, including checks for genotype and allele frequencies, missingness, Hardy–Weinberg equilibrium (p<1×10⁻²⁰), sex concordance and poorly performing probes. SNPs failing quality control will be excluded.

GRS for coronary heart disease and related cardiometabolic traits will be sourced from the Polygenic Score Catalog²³ and computed using R. Imputation of missing variants not captured by the Cardio-MetaboChip will be performed using the Michigan Imputation Server²⁴ or similar. The GRS and allele frequencies in FHS participants will be compared with healthy volunteers and patients with CVD who do not have a family history of premature CVD. Linear and logistic regression will be performed to estimate the individual and cumulative SNPs effect sizes for top-ranked SNPs, adjusted for age, gender and established CVD risk factors as appropriate.

Associations between GRS and clinical variables (e.g., anthropometric traits, biomarkers, outcomes) will be tested using linear regression for continuous traits and logistic regression for categorical traits, adjusting for age, sex and ethnicity. The Cox proportional-hazards model will be used to assess associations between GRS and time-to-event outcomes (e.g., all-cause mortality, cardiovascular death, myocardial infarction and stroke). Univariate analyses (Chi-squared, ANOVA) will be used for exploratory comparisons.

Finally, participants may be screened for monogenic variants associated with CVD. The number of carriers among FHS participants will be compared with subsets from the HVOL and CDCS cohorts.

DNA methylation data will undergo quality control and normalisation in GenomeStudio (Illumina), with batch correction and probe filtering. ANOVA (false discovery rate corrected) will be used to assess differential methylation, and principal component analysis will explore clustering patterns. Associations between methylation, genotype, biomarkers and health outcomes will be tested using linear and logistic regression, Pearson correlations and Chi-squared tests. Pathway analysis will be performed using Ingenuity Pathway Analysis (QIAGEN, Hilden, Germany).  

Results

Recruitment began in January 2010. Of 472 patients screened, 432 (91.5%) were excluded due to ineligibility or lack of contact. Of the 40 eligible individuals, nine declined and three were unable to attend. The current cohort includes 28 participants (Figure 1). Data and sample collection are shown in Figure 2.

Baseline characteristics

The baseline characteristics of the FHS are shown in Table 1. The cohort currently includes equal numbers of males and females, predominantly of European descent (89%). The average age of CVD onset was 46.6 years, and each participant had a median of two first-degree relatives with early-onset CVD. Most (75%) experienced myocardial infarction. Common comorbidities included dyslipidaemia (96%), hypertension (61%) and diabetes (25%). The mean BMI was 28.4kg/m² and 14.3% were current tobacco smokers.

Cardiovascular outcomes

Over a median 11.5-year follow-up, 19 participants had 158 hospital admissions. Eleven individuals accounted for 46 cardiovascular re-admissions, including four for myocardial infarction and 42 for angina.

DNA quality and quantification

DNA samples (mean=104±33ng/µL) had a 260/280 ratio of light absorbance (nm) of ~1.98, indicating high DNA purity. Genotyping was completed for 21 participants and methylation profiling for all 28.

Cardio-MetaboChip SNPs quality control

Of the 196,725 SNPs assayed, 11,211 were excluded due to genotype failure, Hardy–Weinberg deviation or alignment issues. After filtering, 185,514 SNPs (94.3%) remained for analysis (Table 2).

EPIC array DNA methylation quality control

Quality control of 450K methylation arrays retained 454,608 high-quality CpG sites after filtering. All samples passed bisulfite conversion and DNA quantity thresholds. Methylation distributions showed expected bimodal patterns. Sex checks confirmed 100% concordance with documented sex. Blood cell composition was estimated using the Houseman method²⁵ for downstream correction.

Discussion

The FHS will investigate how common and rare genetic variants contribute to premature CVD risk in individuals with premature disease and a family history of CVD. It will provide a valuable resource for assessing relationships between genetic and epigenetic markers, biomarkers and clinical outcomes.

This study is designed to evaluate and compare the burden of known genetic risk markers across cohorts, but numbers are not sufficient to identify novel genetic variants.

A family history of CVD significantly elevates risk, with first-degree relatives conferring 1.4–2.5-fold higher risk, particularly when the onset is premature.²⁶ Risk increases with increased number of affected relatives.²⁷ Yet few studies have examined the cumulative genetic burden in this group,²⁸ and none have assessed GRS in individuals with both personal and familial premature CVD.

In some families, CVD genetic risk arises from rare monogenic variants, and in others from a high burden of common variants. As previously hypothesised, this risk may reflect rare, polygenic or combined genetic factors. By recruiting one individual per family, the FHS is well positioned to explore both monogenic and polygenic contributions.

Recruitment proved challenging. Many potential participants lacked documentation of relatives’ CVD events, particularly when these occurred overseas. Despite significant efforts, only 28 participants were enrolled over 3 years. Non-response, potentially influenced by postal contact methods, may also have introduced selection bias. Additionally, the Cardio-MetaboChip used for genotyping may not include recently discovered CVD-associated variants, even with imputation.

Recruitment to the FHS is currently paused due to funding limitations, although ethical approval remains in place to allow recommencement. The current cohort represents an initial phase of the study and is presented here to describe the data collected to date and to support future funding applications. The small sample size limits statistical power, particularly for sub-group analyses and exploratory investigations. Future work may include expansion of the cohort and integration of multi-omics approaches, subject to funding availability. Early-onset CVD was defined using standard age thresholds, which are largely derived from European populations and may not fully capture earlier disease onset in other ethnic groups. The low number of Māori participants limits the ability to assess genetic risk within Māori. Addressing these limitations would enhance the reliability of study results and provide more comprehensive insights into the molecular pathways underlying genetic susceptibility to premature CVD in New Zealand.

Participants underwent comprehensive clinical and genetic assessment, including biomarker profiling and array-based genotyping and methylation analysis. This enables integrated analysis of genetic risk and its association with CVD traits and outcomes. While clinical scores assess short- to mid-term CVD risk, polygenic scores may offer early, lifelong prediction.²⁹ The FHS will allow investigation of their added value for identifying high-risk individuals with strong family history.

An initial screen for monogenic variants indicated that these variants are poorly represented on the Cardio-MetaboChip. To address this limitation, collaborative work with Canterbury Health Laboratories is planned to genotype monogenic FH variants in our cohort. This approach will enable more comprehensive evaluation of rare inherited variants and allow re-assessment of combined monogenic and polygenic genetic risk in individuals with premature CVD and strong family history.

In summary, the FHS provides a rare and valuable resource for exploring the genetic basis of premature CVD in New Zealanders with strong familial risk, representing an initial cohort to support future expansion and investigation.