In this investigator-initiated study, all researchers were independent of the funding bodies and had complete access to all data. Ethics approval was obtained from the Northern Y Regional Ethics Committee (New Zealand).
During a 14-month period (March 2006 - May 2007), 294 patients with type 2 diabetes diagnosed by their general practitioner, of at least 5 years duration and/or on treatment for type 2 diabetes, were prospectively recruited from primary care. Patients with known cardiac disease (including coronary heart disease, heart failure, LVH (identified on previous ECG or echo done for clinical purposes), moderate and severe valvular heart disease, atrial fibrillation), cerebrovascular disease (prior stroke or transient ischaemic attack), peripheral arterial disease, Stage 3 chronic kidney disease (eGFR < 60 mL/min) or inability to provide informed consent were excluded. A general practitioner (GP) network previously developed in the Natriuretic Peptides in the Community study  was used to facilitate recruitment of primary care patients. Patients meeting study inclusion and exclusion criteria were identified and referred to the study centre by 51 participating GPs within the Auckland region. Study personnel contacted referred patients, provided further details regarding the study and invited them to a study visit.
All patients were seen and evaluated in the Cardiovascular Research Clinic at The University of Auckland. During the study visit patient eligibility was confirmed and informed consent was obtained. Basic demographics and medical history including information regarding known microvascular complications of type 2 diabetes were recorded. The mean of three seated blood pressure (BP) measurements, separated by a minimum interval of five minutes, was obtained. Height, body mass, waist circumference, hip circumference, and body composition were measured. Blood was collected using standard venepuncture technique and samples were sent to a tertiary referral medical laboratory for measurement of creatinine, glucose, HbA1c, lipids, and NT-proBNP. A single urinary albumin:creatinine ratio measured within 12 months of the study visit was obtained from community laboratories. If this was unavailable, participants were directed to have this done soon after the study visit.
All patients had a resting transthoracic echocardiogram (Philips HDI 5000/iE33, Bothell, Seattle, Washington) which was the reference standard for the detection of LVH in this study performed by a research-trained sonographer. LV mass was assessed from M-mode images in accordance with The American Society of Echocardiography (ASE) guidelines . When M-mode images were unsuitable for measurement, 2-dimensional images were used. The LV mass gender-specific cut-offs for LVH used were: >45 g/m2.7 in women and >49 g/m2.7 in men. All echocardiographic measurements were made by a cardiologist with subspecialty training in echocardiography. For LV mass the coefficient of variability for intra-observed repeated measures is less than 8% . The echocardiographer and cardiologist measuring the images were blinded to ECG and NT-proBNP results.
All patients had a standard unfiltered 12-lead ECG (Philips Hewlett-Packard PageWriter 200 Cardiograph, Andover, Massachusetts). Each ECG was measured by a single analyst using a 150 mm digital vernier calliper under a five-fold magnification. ECG criteria used for the detection of LVH included the Sokolow-Lyon (SV1 + RV5/6) and Cornell (women RaVL + SV3 + 0.8 mV; men RaVL + SV3) voltage criteria. Standard cut-offs for the electrocardiographic diagnosis of LVH were used: Sokolow-Lyon voltage >3.5 mV and Cornell voltage >2.8 mV. Patients meeting either criterion were considered to have LVH by ECG. The ECG analyst was blinded to echocardiographic and NT-proBNP results. Thirty participants were randomly selected for estimation of test reproducibility. This analysis demonstrated an intra-observer and inter-observer variability of ≤0.01 mV for measurement of ECG voltages.
NT-proBNP levels were measured using the Roche Diagnostics Elecsys assay (pmol/L). The performance characteristics claimed by the manufacturer are an analytical sensitivity of 0.6 pmol/L and functional sensitivity of <5.9 pmol/L.