Singapore is a developed city state of 5.6 million people in Asia with well-represented diverse ethnicities including Chinese (74.3%), Malays (13.4%) and Indians (9.1%). The elderly ≥65 yr account for 12.4% of the population and the average life-expectancy is 82.9 years old .
The SingHEART study is the first multi-ethnic prospective population-based study of healthy Asians harnessing the latest technologies. In summary, healthy male and female subjects aged 21–69 years old without any prior cardiovascular disease (Ischemic heart disease, stroke, peripheral vascular disease) or diabetes mellitus will be recruited from the general population. The complete inclusion and exclusion criteria are found in the supplementary appendix (Additional file 1: Appendix 1). These patients are drawn from a cohort of healthy individuals who have already volunteered to the institutional Biobank program. Subjects fulfilling the inclusion criteria will be recruited from the public via advertisements (e.g. posters, local newspaper).
Written informed consent will be obtained from every subject and includes follow-up of up to 20 years, including agreement to permit tracking of outcomes via national and disease registries. Subjects will be informed of incidental abnormalities picked up during the tests. If genetic abnormalities are detected, there is an option for genetic counselling. All data will be anonymized for analysis.
Ethical approval was obtained from the Singhealth institutional review board.
This study aims to comprehensively characterize a healthy Asian population at baseline and at follow-up, using multiple modalities to aid in the understanding of the traditional and novel factors that contribute to the development of cardiovascular disease. Emphasis will be placed on understanding how these traditional and novel factors interact, as well as on elucidating the ethnic differences in the development of cardiovascular disease. The specific objectives are as follows:
To characterize cardiovascular health in Asia by measuring multiple systems simultaneously and longitudinally.
To assess lifestyle, diet, physical activity and sleep via traditional questionnaire surveys and wearable technologies and their impact on the development of cardiovascular disease.
To characterize baseline genetic, metabolic and advanced cardiac imaging profiles in healthy individuals and the identification of novel markers influencing the development of cardiovascular disease with the potential for therapeutics.
Validate patient-reported sleep/physical activity and that derived by wearables, study the impact of physical activity on cardiac structure, investigate relationship between calcium score and lifestyle factors, etc.
To study the development and progression of cardiovascular disease in these patients to help develop new preventive, diagnostic and predictive tools for the Asian population
To understand the differential effects of ethnicity on the development of cardiovascular disease
To use both traditional statistical analysis and newer methodologies (e.g. machine learning) to process and analyze the data
Each subject will undergo the following investigations at baseline and at specific intervals on follow-up for up to 20 years, as described in Table 1.
The questionnaire will include sections on demographics, socioeconomic status, medical history, lifestyle, diet and exercise, quality of life as assessed by the EQ-5D-3 L, Pittsburgh sleep quality index and International Index of Erectile Function (IIEF)- 5 (only for males). See supplementary appendix for the complete questionnaire (Additional file 2: Appendix 2).
Basic blood investigations
This will include full blood count, renal and liver function, fasting lipids and fasting glucose. As diabetes mellitus is an exclusion criterion, HbA1c will not be measured routinely. Samples are biobanked for future use through the NHCS biobank.
A standard 12-lead resting ECG will be recorded. Variables studied will include rate, rhythm, axis, conduction intervals, morphologies (including QRS, S-T and T wave abnormalities) and arrhythmias.
Ambulatory BP monitor
Ambulatory blood pressure monitoring will be performed for 24 h via a cuff-monitoring (Spacelab Healthcare Model 90,227/90217A). Data collected will include the systolic, diastolic and mean blood pressure during the day and night, as well as data on dipping.
Continuous ECG monitoring
Continuous ECG monitoring will be performed for 24 h–72 h via a wearable, multi-lead ECG monitoring patch which stores data for up to 3 days (ePatch® ECG recorder AMS3000). Data collected will include variations in heart rate and the various types of arrhythmias present.
Activity and sleep tracker
A commercially available wearable fitness device able to track heart rate, physical activity, exercise and sleep will be used. The participants will wear this activity tracker for 5–7 days. Currently, the study uses the Fitbit Charge HR. Data for each subject will be downloaded from the Fitbit Application Programming Interface (https://dev.fitbit.com/reference/web-api/quickstart). Step counts are available at two levels; intraday step counts in 15-min intervals and daily totals. Intraday HR data are available at 5-min intervals, along with confidence levels. Intraday sleep tracking data containing details of each sleep session will also be recorded.
Electron beam CT scanning using contiguous 3-mm slices during a single breath hold will be performed by a 320-slice CT scanner (Toshiba Aquilion ONE) with the following parameters: tube voltage 120kVp, tube current 200-400 mA (dependent on patient size and shape as visually assessed by the radiographers), gantry rotation time 350 ms and 3 mm section thickness. The non-contrast scan will be volume prospective and ECG-gated and synchronized to the RR interval with a scan time of 100 ms/slice. A calcified lesion will be defined as at least two contiguous voxels with an attenuation coefficient > 130 Hounsfield units. Coronary calcium scores will be calculated as previously described  via Vitrea Workstation.
Cardiovascular phenotyping using cardiovascular magnetic resonance (CMR) will be performed in all healthy volunteers (3 T Ingenia, Philips or 1.5 T MAGNETOM Aera, Siemens). Conventional balanced steady-state free precession cine images of the vertical and horizontal long-axis planes and the sagittal LV outflow tract view will be acquired. Short-axis cines will be obtained from the mitral valve annulus to the apex (8 mm slices with 2 mm gap). In addition, a single breath hold 3D LV short axis stack will also be acquired in the same orientation. Aortic flow will be assessed using velocity-coded phase contrast imaging. Cardiac volumes, left ventricular mass, atrial sizes and aortic root will be measured in all patients using the CMR 42 software (Circle Cardiovascular Imaging). Normal CMR reference ranges in the Asian population were recently published using standardized protocols in our Image Analysis Laboratory .
20 μl of lipid internal standard mix and 10 μl of 14:0 phosphatidylcholine will be added to 100 μl of serum in a microcentrifuge tube. After an equilibration period of 30 s, 1.2 ml of HPLC-grade methanol will be added to the mixture, followed by vortexing. The mixture will be incubated at 50 °C for 10 min, followed by centrifugation to pellet the precipitated protein. The supernatant will then be removed and placed in a clean microcentrifuge tube for drying under nitrogen gas. 100 μl of methanol will be used to reconstitute the dried extract. The reconstituted lipid solution will be separated using a LC-MS (liquid chromatography – mass spectrometry) system (Agilent 1260) and a Thermo Scientific Accucore HILIC column (100 × 2.1 mm; particle size 2.6 μm). Mobile phase A consists of acetonitrile/water (95:5) with 10 mM ammonium acetate, pH 8.0 and mobile phase B consists of acetonitrile/water (50:50) with 10 mM ammonium acetate, pH 8.0. For the separation, the column will be equilibrated with 100% mobile phase A, increasing to 20% mobile phase B in 5 min, then held for 5 min. The column will finally be re-equilibrated with 100% mobile phase A for 5 min. Finally, mass spectrometry and data acquisition will be performed using an Agilent 6430 triple-quadrupole mass spectrometer .
Whole genome sequencing will be done in Illumina HiSeq X sequencers at 30X coverage. 1.0 micrograms of DNA per sample will be used for library preparation. Sequencing libraries will be generated using the Truseq Nano DNA HT sample preparation kit (Illumina, SUA) and idenx codes will be added to each sample. Briefly, the genome will be fragmented by sonication to a size of 350 bp, and the DNA fragments will be end-polished, A-tailed and ligated with full-length adapter for Illumina sequencing with further PCR amplification. Lastly, PCR products will be purified (AMPure XP system) and the libraries analysed for size distribution by Agilent 2100 Bioanalyzer and quantified using real-time PCR.
Outcomes will be tracked 6 monthly for 20 years via review of medical records. The outcomes studied will include mortality and cause of death, myocardial infarction, stroke, malignancy, heart failure, and the development of comorbidities (e.g. ischemic heart disease, peripheral vascular disease, diabetes mellitus, renal failure, etc.). Patients provided explicit consent for the matching of outcomes against databases and registries. Information on mortality, myocardial infarction, stroke, chronic kidney disease and malignancy will be obtained from national state-mandated registries (e.g. Singapore Myocardial Infarction Registry, Singapore Stroke Registry, Singapore Renal Registry, Singapore Cancer Registry, Singapore Renal Registry, etc).
The SingHEART Steering Committee is responsible for the overall conduct of the study. The protocol design and execution of the study is entirely under the oversight of the SingHEART Steering Committee, and the funding sources have no access to patient-specific data. Because volunteers are drawn from the Biobank cohort, the SingHEART program adheres to protocols managing patient data anonymization, patient confidentiality/privacy, incidental finding management; and biosample/genomic DNA samples. Data is entered electronically into a pre-designed database by trained study coordinators and the accuracy of the data will be regularly audited. Requests to use the data or biospecimens for research will require approval from the steering committee in accordance with standardized procedures. The study team meets once every month to review study progress and to address any concerns raised by the study coordinators.
The initial target enrolment is 2500 based on feasibility and initial funding availability, with the opportunity to extend recruitment with further funding. Multiple linear regression and logistic regression analyses using generalised linear models will be performed to study the relationships between factors. Odds ratios, 95% confidence intervals and p-values will be reported. Cox-proportional hazards models will be used to analyse longer-term health outcomes and hazards ratios reported. Multivariate adjustment for confounders will be performed. Machine learning methodology appropriate for high-dimensionality datasets (e.g. deep learning systems and neural networks, classification techniques involving decision trees and probabilistic prediction) will also be performed in tandem with conventional statistical analysis.