There is considerable evidence to show that obesity is associated with the development of heart failure independent of traditional risk factors. However, clarity is required on the process involved in the transition from obesity-associated left ventricular hypertrophy (LVH) to LV dysfunction. In the present thesis I evaluated the extent to which central obesity explains variations in LV diastolic function at a community level independent of LV mass (LVM), LV remodelling or haemodynamic factors; whether obesity-related increases in LVM exceeding that predicted by workload (inappropriate LVM [LVMinappr] or alternative haemodynamic factors explains variations in LV ejection fraction (EF) at a community level; whether regression of LVMinappr is more closely associated with improvements in EF than LVM or LVM index (LVMI); and whether obesity-associated insulin resistance may explain decreases in LV diastolic function and variations in LVMinappr. Data were obtained in either 626 or 478 participants whom were representative of a randomly selected community sample and in 168 mild to moderate hypertensives treated for 4 months. In 626 randomly selected participants over 16 years of age from a community sample with a high prevalence of excess adiposity (~24% overweight and ~43% obese) after adjustments for a number of confounders including age, sex, pulse rate, conventional diastolic (or systolic) blood pressure (BP), antihypertensive treatment, LVMI and the presence of diabetes mellitus or an HbA1c>6.1%; waist circumference (p=0.0012) was independently and inversely associated with a reduced early-ro-late transmitral velocity (E/A), with similar findings noted for e’/a’ in a subset of 212 participants with tissue Doppler measurements. Waist circumference-E/A relationships persisted even after adjustments for other adiposity indices including body mass index (BMI) (p<0.05-0.005). No independent relationships between adiposity indices and E/e’ were noted (n=212). In contrast to the effects on diastolic function, waist circumference was not correlated with EF (p=0.83). The independent relationship between waist circumference and E/A was second only to age and similar to BP in the magnitude of the independent effect on E/A. The inclusion of relative wall thickness rather than LVMI in the regression equation produced similar outcomes. The inclusion of carotid-femoral pulse wave velocity (PWV), or 24-hour BP as confounders, failed to modify the relationship between waist circumference and E/A. Thus, waist circumference is second only to age in the impact of the independent association with E/A in a community sample with a high prevalence of excess adiposity. This effect was not accounted for by left ventricular hypertrophy or remodelling, 24-hour BP or arterial stiffness. In 478 randomly selected participants from a community sample, waist circumference, but not BMI was independently associated with the homeostasis model assessment of insulin resistance (HOMA-IR). HOMA-IR was inversely correlated with E/A (p<0.0001) and in a multivariate model with adjustments for waist circumference, age, sex, conventional diastolic or systolic BP, diabetes mellitus or an HbA1c>6.1%, regular tobacco use, regular alcohol intake, pulse rate, treatment for hypertension and either LVMI or LV relative wall thickness in the model, the relationship betwreen HOMA-IR and E/A persisted (partial r=-0.13 to 0.14, p<0.005). With further adjustments for either 24-hour systolic or diastolic BP (partial r=-0.11, p<0.05, n=351) or for aortic PWV (partial r=-0.11, p<0.02, n=410), the independent relationship between HOMA-IR and E/A also remained. Therefore, the relationship between indices of an excess adiposity and abnormalities in LV diastolic function may be explained in-part by insulin resistance beyond haemodynamic factors. In 626 randomly selected adult participants from a community sample with a high prevalence of obesity, the strongest independent predictor of LVMinappr was BMI (p<0.0001). With adjustments for LV stress and other confounders there was a strong inverse relationship between LVMinappr and EF (partial r=-0.41, p<0.0001), whilst only modest inverse relations between LVM or LVMI and EF were noted (partial r=-0.07 to -0.09, p<0.05-0.09)(p<0.0001, comparison of partial r values). The independent relationship between LVMinappr and EF persisted with further adjustments for LVM or LVMI (partial r=-0.52, p<0.0001). LVMinappr and LV midwall fractional shortening were similarly inversely related (p<0.0001) and these relations were also stronger than and independent of LVM or LVMI. In conclusion, in a community sample with a high prevalence of obesity, inappropriate LVM is strongly and inversely related to variations in EF independent of and more closely than LVM or LVMI and BMI was the strongest independent determinant of inappropriate LVH. Therefore LVH is a compensatory response to workload, but when exceeding that predicted by workload, as may occur in obesity, is associated with LV systolic chamber decompensation. In 168 mild-to-moderate hypertensives treated for 4 months, although in patients with an LVMI>51g/m2.7 (n=112)(change in LVMI=-13.7±14.0 g/m2.7, p<0.0001), but not in patients with an LVMI≤51g/m2.7(n=56)(change in LVMI=1.3±9.3 g/m2.7) LVMI decreased with treatment; treatment failed to increase EF in either group (1.2±10.8% and 2.7±10.7% respectively). In contrast, in patients with inappropriate LVH (LVMinappr>150%, n=33) LVMinappr decreased (-32±27%, p<0.0001) and EF increased (5.0±10.3%, p<0.0001) after treatment, whilst in patients with a LVMinappr≤150% (n=135), neither LVMinappr (-0.5±23%), nor EF (0.9±10.3%) changed with therapy. With adjustments for circumferential LV wall stress and other confounders, whilst on-treatment decreases in LVM or LVMI were weakly related to an attenuated EF (partial r=0.17, p<0.05), on-treatment decreases in LVMinappr were strongly related to increases in EF even after further adjustments for LVM or LVMI (partial r=-0.63, confidence interval=-0.71 to -0.52, p<0.0001). In conclusion, decreases in LVMinappr are strongly related to on-treatment increases in EF beyond changes in LVM and LVMI. LVH can therefore be viewed as a compensatory change that preserves EF, but when in excess of that predicted by stroke work, as a pathophysiological process accounting for a reduced EF. In 478 participants of a randomly selected community sample with adjustments for waist circumference, age, sex, conventional systolic BP, diabetes mellitus or an HbA1c>6.1%, regular tobacco use, regular alcohol intake, pulse rate, and treatment for hypertension, an independent relationship between HOMA-IR and LVMinappr was noted (partial r=0.14, p<0.002). With further adjustments for either 24-hour systolic BP (partial r=0.11, p<0.05, n=351), aortic PWV (partial r=0.13, p<0.02, n=410), or circumferential LV wall stress (partial r=0.12, p<0.02, n=478) the independent relationship between HOMA-IR and LVMinappr also remained. Thus, the relationship between indices of an excess adiposity and LVM beyond haemodynamic factors may be explained in-part by insulin resistance. In conclusion, the results of the present thesis provide clarity on the process involved in the transition from obesity-associated LVH to LV dysfunction. In the present thesis I demonstrated that an index of central obesity explains a considerable proportion of the variation in LV diastolic function at a community level independent of LVM, LV remodelling and haemodynamic factors; that obesity-related increases in LVM exceeding that predicted by workload (LVMinappr) or alternative haemodynamic factors explains a marked proportion of variations in EF at a community level; that regression of LVMinappr is more closely associated with improvements in EF than LVM or LVM index (LVMI); and that obesity-associated insulin resistance may explain decreases in LV diastolic function and variations in LVMinappr and hence EF. Therefore, studies are warranted to evaluate the impact of interventions that improve insulin sensitivity on obesity-related decreases in LV diastolic function and increases in LVMinappr.