Discussion
This study demonstrated that in individuals with non-ischaemic stage A HF risks, T2DM is associated with more impaired cardiac function and reduced exercise capacity than is present in those with hypertension. Although patients with well-controlled and poorly controlled BP showed abnormal diastolic function, it appears that abnormal GLS is an independent marker for diabetic cardiomyopathy rather than hypertensive heart disease. Poor BP control is associated with more impaired cardiac function with or without the presence of diabetes.
Combined effect of T2DM and hypertension on LV function
Diabetes and hypertension constitute two powerful independent risk factors for cardiovascular disease. T2DM is known to be a strong predictor of incident HF, independent of other concomitant risk factors.6 ,17–21 Subclinical diastolic dysfunction and systolic impairment assessed using GLS are believed to be early markers of diabetic cardiomyopathy.5 ,18 However, hypertension is present in 40–80% of patients with long-standing diabetes,22 and most of these studies were performed in populations with a high prevalence of hypertension and therefore reflect the combined impact of hypertension and T2DM. In our study, patients with mixed T2DM and hypertension had a 20% prevalence of E/e′>15, analogous to a 23% prevalence in another community-based study of 1760 patients with T2DM with 86% of hypertension and 36% prevalence of CAD. Follow-up of that group showed that the HR of hypertension (HR 4.27, 95% CI 1.92 to 12.15) for subsequent HF was almost double that of CAD (HR 2.2, 95% CI 1.62 to 3.01). The negative synergistic effect of hypertension and diabetes was likely the cause of high prevalence of impaired diastolic and systolic dysfunction and associated adverse outcome.6 ,18 ,23 However, the exact underlying pathophysiology of this combined impact is unclear. Diabetes is a metabolic disorder characterised by intracellular accumulation of toxic fatty acid intermediates.24 This change also affects cardiac mitochondria, resulting in contractile dysfunction.25 There is a well-recognised tendency to develop diastolic dysfunction even in the absence of significant hypertension; however, the presence of hypertension may accelerate the adverse changes and cause end-organ damage.26 Quantitative measure using fibrosis score showed the degree of myocardial and interstitial fibrosis contributes to the pathological involvement.27 The score was found to be lowest for hypertensive, midrange for diabetic and highest for hypertensive diabetic. It is presumed that fibrosis and metabolic consequences of myocyte in diabetes lead to impaired systolic and diastolic function, while chronic afterload causes interstitial fibrosis, leading to a more impaired diastolic than systolic function in hypertension. The coexisting hypertension exacerbates functional changes by producing larger amount of fibrosis. Another observation was described that abnormal GLS and diastolic dysfunction were not analogous to each other. As an early marker, diastolic function was documented in 47% of patients with T2DM, Ernande showed abnormal strain in 28% of those with normal diastolic function.5 In multivariable analysis, a history of hypertension but not T2DM was associated with diastolic parameters. This relationship was mirrored in our study, in which the prevalence of diastolic dysfunction was 72% in those with T2DM with abnormal strain in 47% of them (table 2)—a higher prevalence found in our study was likely due to older age (71±5 vs 52±5 years) and higher prevalence of history of hypertension (67% vs 38%). A history of hypertension but controlled BP was associated with increased E/e′, which may represent a combined impact. The findings parallel the finding that hypertension (either historical or high BP at the time of the echocardiogram) was independently associated with e′ and E/e′ and diabetes was associated with E/e′.23
It needs to be noted that our finding of GLS consistently associated with diabetes but not hypertension in the multivariable analysis should not be interpreted as a normal GLS in this population. Influence of afterload on LV causing reduced GLS in early disease stage was described in animal model and human studies.28–30 Understanding these differences would be important and beneficial to guide effective screening and early intervention in the community as hypertension and diabetes are the two leading aetiologies of preclinical HF in this population.
Effects of controlled and uncontrolled hypertension on LV impairment
Hypertension has been shown to precede the development of HF in men and women.31 Although there have been improvements in the overall management of hypertension, there remain a significant number of hypertensive patients who remain untreated or fail to achieve optimal control.32 ,33 Of the 82% with a known history of hypertension in our study, 92% were on antihypertensive therapy, but only 33% had good control of BP (table 1). Our study demonstrated uncontrolled BP was independently associated with more severe cardiac dysfunction including abnormal e′, E/e′, DS, DSR and LV mass. However, GLS appeared to be relatively preserved in those with hypertension compared with those with neither hypertension nor T2DM. These findings are inconsistent with previous work in a small group of younger (46±14 years) hypertensive patients with controlled BP showing lower peak strain and strain rate at rest, with blunting of strain increment during exercise.29 The dependence of myocardial strain on haemodynamic conditions has been reported in hypertension34 ,35 and valve disease.36
Assessment of exercise capacity using 6MW
Impaired exercise capacity and functional changes during exercise were known to be early markers of subclinical LV dysfunction in patients with hypertension and diabetes.37–39 However, a standard exercise testing protocol is not feasible in community-based screening for subclinical LV dysfunction. Owing to its simplicity and inexpensiveness, the 6MW test is often used to estimate submaximal functional capacity in this setting; the predictive value of 6MW for peak oxygen uptake is of moderate accuracy.40 In our study, 6MW distance correlated with subclinical cardiac dysfunction and was significantly reduced in those with T2DM+HTN+ individuals but relatively preserved in those with hypertension alone.
Limitations
The present analysis was based on a cross-sectional sample from a clinical trial population of participants aged ≥65 years with at least one of the listed non-ischaemic stage A HF risks. The control group without T2DM or hypertension had other HF risks (mainly obesity), but there were no age-matched controls without HF risk factors. Another important limitation of this study was the concomitant presence of CAD was not assessed. Our intention and focus was on non-ischaemic population with a very low prevalence of known CAD (<5%). However, diabetic cardiomyopathy and hypertensive heart disease are known as part of atherosclerosis process, which make their heart susceptible to ischaemia coronary changes. Some of the functional change may be caused by underlying ischaemic and non-ischaemic pathophysiological changes. A possible approach to address this limitation would be a stress test to identify those with underlying CAD, but we could not perform this in the context of a community-based study.
Conclusions
Hypertension is associated with less impairment of GLS and exercise capacity than is T2DM. Those with well-controlled and poorly controlled BP showed abnormal diastolic functional markers, and more severely impaired cardiac function was associated with worse BP control. However, GLS appears to be associated with diabetic cardiomyopathy rather than hypertensive heart disease in this population at risk of HF.