Sex-specific pattern of left ventricular hypertrophy and diastolic function in patients with type 2 diabetes mellitus

Abstract

Aims 

Few prospective studies have evaluated sex-specific pattern, natural progression of left ventricular (LV) remodelling, and diastolic dysfunction in patients with type 2 diabetes (T2DM). The aim of this study was to study the sex-specific prevalence, longitudinal changes of LV remodelling, and diastolic dysfunction in patients with T2DM. Further, the prognostic value of diastolic function in women and men was also evaluated.

Methods and results 

A total of 350 patients with T2DM (mean age 61 ± 11 years; women, 48.3%) was recruited. Detailed echocardiography was performed at baseline and after 25 months. A major adverse cardiovascular event (MACE) was defined as cardiovascular death, heart failure hospitalization, or myocardial infarction. Despite a similar age, prevalence of hypertension and body mass index, women had a higher prevalence of LV hypertrophy and diastolic dysfunction at baseline and follow-up compared with men. A total of 21 patients developed MACE (5 cardiovascular death, 9 hospitalization for heart failure, and 7 myocardial infarction) during a median follow-up of 56 months. Women with diastolic dysfunction had a higher incidence of MACE than those with normal diastolic function but this association was neutral in men. Multivariable Cox-regression analysis indicated that diastolic dysfunction was associated with MACE in women [hazard ratio = 6.30; 95% confidence interval (CI) = 1.06–37.54; P < 0.05] but not men (hazard ratio = 2.29, 95% CI = 0.67–7.89; P = 0.19).

Conclusion 

LV hypertrophy and diastolic dysfunction, both at baseline and follow-up, were more common in women than men. Pre-clinical diastolic dysfunction was independently associated with MACE only in women with T2DM but was neutral in men.

Introduction

Type 2 diabetes mellitus (T2DM) is associated with left ventricular hypertrophy (LVH) and diastolic dysfunction, independent of underlying coronary artery disease.¹ Studies have consistently demonstrated that the presence of adverse left ventricular (LV) remodelling and diastolic dysfunction is common in patients with T2DM, even before the development of symptoms.^2–4 In particular, the presence of diastolic dysfunction in patients with T2DM has been shown to be associated with a 1.62-fold risk of future heart failure and 2.18-fold risk of mortality compared with those with normal diastolic function.⁵

There is increasing evidence that sex-specific differences are important in the epidemiology, pathophysiology, and outcome of cardiovascular diseases.^6,7 Women with heart failure and preserved ejection fraction (HFpEF) exhibit a greater level of LVH and diastolic dysfunction than men.⁸ There are several cross-sectional echocardiography studies that consistently confirm that diabetic women has a distinct type of LV remodelling⁹ and diastolic dysfunction¹⁰ compared with diabetic men. However, there are limited data that evaluate the changes of sex-specific pattern of LV remodelling and diastolic dysfunction by prospective echocardiography studies. This study aimed to evaluate sex-specific prevalence, longitudinal changes of LV remodelling, and diastolic dysfunction in patients with T2DM. We further investigated differences in the prognostic value of diastolic dysfunction in women and men.

Methods

Study population

A total of 382 consecutive patients with T2DM were recruited at Queen Mary Hospital and the University of Hong Kong Shenzhen Hospital from January 2012 to April 2017. The diagnosis and classification of T2DM were according to the World Health Organization criteria.¹¹ Patients with a history or clinical symptoms and signs suggestive of cardiovascular disease including severe valvular heart disease (n = 1), heart failure (n = 5), atrial fibrillation (n = 4), history of coronary artery disease (n = 6), malignancy (n = 3), end-stage renal failure (n = 7), or poor image (n = 6) were excluded. A total of 350 patients were recruited for baseline evaluation. Patients were subsequently excluded from follow-up echocardiography if they had developed new-onset atrial fibrillation (n = 1), heart failure (n = 3), myocardial infarction (n = 1), death (n = 1), malignancy (n = 5), end-stage renal failure (n = 6), were lost to follow-up (n = 21), or declined the invitation (n = 16). The number of patients recruited for follow-up echocardiographic study was 296. The study was approved by the ethics committee of the West Cluster Hospital Authority of Hong Kong and all subjects gave written informed consent. This study was part of the Chinese Diabetic Heart Study (CDATS) to evaluate cardiovascular manifestations in Chinese patients with T2DM that attempted to evaluate the pathophysiology and potential therapies.¹²

Clinical data

Data on baseline and follow-up clinical characteristics, blood sampling, and clinical assessment were obtained on the same day for all study subjects following overnight fasting for at least eight hours. Conventional cardiovascular risk factors such as history of smoking, hypertension, and hypercholesterolaemia were documented. Blood pressure was measured at the end of echocardiography after a 5-min rest. Pulse pressure (PP) was defined as the difference between systolic and diastolic blood pressure. Height and weight were measured, and body surface area (BSA) and body mass index (BMI) calculated according to the appropriate formula.

Major adverse cardiovascular events

A major adverse cardiovascular event (MACE) was defined as cardiovascular death, hospitalization for heart failure, or myocardial infarction. Patients were censored at the first event: multiple events were not counted. Follow-up evaluation at the end was August 2018. All clinical endpoints were collected from the inter-hospital electronic clinical management system.

Echocardiography

All echocardiography was performed using a commercially available ultrasound system (Vingmed Vivid 7 or E9, General Electric Vingmed Ultrasound, Milwaukee, WI, USA) by two experienced cardiologists. A 3.5 MHz probe was used to obtained images that were stored in cine-loop format (three cardiac cycles). Two-dimensional guilded M mode echocardiography was used to measure inter-ventricular septum (IVSd), LV dimension (LVDd), and posterior wall thickness (LVPWd) at end-diastole in the parasternal long-axis view according to the guideline of the European Association of Cardiovascular Imaging (EACVI).¹³ Relative wall thickness was calculated according to the formula: RWT = 2×LVPWd/LVDd; LV mass index (LVMi) was calculated according to the Devereux formula and index to BSA.¹⁴ LV geometry was classified based on LVMi and RWT.¹³ LV hypertrophy was defined as LVMi >115 g/m² in men and >95 g/m² in women, and increased RWT as RWT >0.42 in both men and women. Normal LV geometry was classified as normal LVMi and normal RWT. Concentric remodelling was defined as normal LVMi and increased RWT. Eccentric LVH was defined as LVH and normal RWT. Concentric hypertrophy was defined as LVH and increased RWT. LV ejection fraction (LVEF), LV end-diastolic volume (LVEDV), and end-systolic volume (LVESV) were derived from the standard apical four-chamber and two-chamber view using the modified biplane Simpson method. Left atrial volume index (LAVi) was assessed in the apical four-chamber view and indexed to BSA. Peak mitral inflow velocity at early (E) and late diastole (A) was measured by pulsed-wave Doppler with sample volume put at the tips of mitral leaflets. Early diastolic tissue velocity of mitral annular was measured in the septal (e’ septal) and lateral (e’ lateral) basal regions in the apical four-chamber view using tissue Doppler imaging and average E/e’ calculated. Peak velocity of tricuspid regurgitation (TR) was assessed by continuous wave Doppler. Diastolic dysfunction grade was classified according to the 2016 recommendations of the ASE/EACVI guideline by two steps.¹⁵ In the first step, patients were defined as normal, indeterminate, or having diastolic dysfunction according to septal e’ velocity or lateral e’ velocity, average E/e’, LAVi, and peak TR velocity. In the second step, patients with diastolic dysfunction in step one were graded according to E/A ratio and E velocity followed by average E/e’, LAVi, and peak TR velocity.

Statistical analysis

Continuous variables with a normal distribution are presented as mean ± standard deviation. Comparison of normally distributed variables at baseline and follow-up was made by paired t-test and the difference between women and men at baseline was evaluated by independent sample t-test. The skewed variables, including serum creatinine and triglyceride levels, are expressed as median (interquartile range). Non-parametric test was used to compare the difference of serum creatinine and triglyceride level between women and men. Analysis of covariance (ANCOVA) adjusted for age, hypertension, and the duration between two echocardiography was used to evaluate the difference of echocardiography parameters between women and men at follow-up. Further, a linear mixed model adjusted for age, hypertension, LV echocardiography parameters at baseline and duration between two echocardiography was performed to evaluate the association between the change of LV parameters and sex. Categorical variables are presented as frequencies and percentages and were compared using the χ² test. Kaplan–Meier curves were constructed and MACE of diastolic function and sex were compared using the log-rank test. Multivariable cox-regression adjusted for age, PP, HbA1c, and antidiabetic treatment was used to assess the association between sex-specific diastolic function and the development of a MACE. Inter-observer and intra-observer variability were assessed by using Bland–Altman and intraclass correlation coefficient. All statistical analyses were performed using SPSS software (Version 23.0, Chicago, IL, USA). A two-sided P value <0.05 for all tests was considered statistically significant.

Results

Clinical characteristics

A total of 350 patients (48.3% of women) were recruited; their baseline clinical characteristics are shown in Table 1. Age, BMI, and duration of diabetes were similar for women and men. Nonetheless men had a higher diastolic blood pressure and serum creatinine level and were more likely to be a smoker than women. Interestingly, PP was significant higher in women compared with men (Table 1). Women had a higher total cholesterol and high-density lipoprotein cholesterol level. Serum fasting glucose, HbA1c, and the use of medication were similar for women and men. The comparison of baseline and follow-up blood pressure, HbA1c%, glucose and use of medication is shown in Supplementary data online, Table S1.

Baseline echocardiography parameters

Sex-specific echocardiographic parameters at baseline are shown in Table 2. Women had a smaller IVSd, LVPWd, LVMi, LVEDV, and LVESV than men and a higher LVEF. Nonetheless, LVH (defined as eccentric hypertrophy and concentric hypertrophy) was more common in women than men (37.9% vs. 24.8, P < 0.05; Figure 1A). Moreover, diastolic function in women was more impaired, as reflected by a lower e’ septal, lower e’ lateral, higher average E/e’, and larger LAVi. As a result, women had a higher prevalence of diastolic dysfunction ≥ Grade 1 than men (40.2% vs. 27.1%, P < 0.05; Figure 1B).

Echocardiographic parameters at follow-up

The median duration between baseline and follow-up echocardiography for the entire cohort was 25 months (range 12–54 months) and was similar between women and men [25 (range 12–50) vs. 25 (range 12–54) months, P = 0.84]. The entire cohort showed a significant increase in LV structure including IVSd, LVPWd, RWT, LVMi, and LAVi at follow-up, while LVEDV and LVESV remained similar (Supplementary data online, Table S1). In addition, the prevalence of LVH increased from 28.3% to 33.1% of which the proportion of concentric hypertrophy subtype increased from 15.9% to 23.6%. In the entire cohort, systolic function and diastolic function parameters deteriorated at follow-up (Supplementary data online, Table S1). Further, the prevalence of LV diastolic dysfunction was significantly increased (33.4–45.9%) during follow-up compared with baseline echocardiography. The sex-specific echocardiography parameters at baseline and follow-up are summarized in Table 3. IVSd, LVPWd, and LVMi were higher in men compared with women at baseline and follow-up. The prevalence of LVH in women remained higher than in men at follow-up (41.3% vs. 25.9%, P < 0.01) (Figure 2A), as did average E/e’ and the prevalence of diastolic dysfunction (53.6% vs. 39.2%, P < 0.05) (Figure 2B). The linear mixed model showed that women was negatively associated with change in LV wall thickness and LVESV and positively associated with change of E velocity (Table 4). Nonetheless, sex was not associated with the change of LVMi, LVEF, and E/e’ ratio.

Clinical outcomes

The median follow-up was 56 months. A total of 21 patients developed a MACE, 5 cardiovascular death, 9 heart failure hospitalization, and 7 myocardial infarction. The prevalence of MACE between women and men was similar (5.3% vs. 6.6%, P = 0.61). Kaplan–Meier survival curves demonstrated that women and men were at similar risk for adverse events (Figure 3). Patients with diastolic dysfunction had a higher risk for MACE compared with those with normal diastolic function (Figure 4). Sex-specific Kaplan–Meier survival curve demonstrated that only women, not men, with diastolic dysfunction had a higher risk of MACE compared with those with normal diastolic function (Figure 5).

Univariate Cox-regression demonstrated that age was not associated with MACE for the entire cohort [hazard ratio = 1.03, 95% confidence interval (CI) = 0.98–1.07; P = 0.23], women (hazard ratio = 1.05, 95% CI = 0.98–1.12; P = 0.17) and men (hazard ratio = 1.01, 95% CI = 0.96–1.07; P = 0.65). Further, univariate cox regression showed that PP was not associated with MACE in the entire cohort (hazard ratio = 1.02, 95% CI = 0.99–1.04; P = 0.17), women (hazard ratio = 1.02, 95% CI = 0.99–1.05; P = 0.19) nor men (hazard ratio = 1.01, 95% CI = 0.98–1.05; P = 0.41). Unadjusted Cox-regression demonstrated that the presence of diastolic dysfunction was predictive of MACE in women but did not reach statistical significance in men. Multivariable Cox regression analysis adjusted for age, PP, HbA1c, and antidiabetic treatment further demonstrated that diastolic dysfunction was associated with MACE only in women, not men (Table 5).

Intraobserver and interobserver variability

Intraobserver and interobserver variability for echocardiography parameters are shown in Table 6. All variability was within acceptable limits.

Discussion

This study provides a prospective evaluation of the sex-specific prevalence and longitudinal changes to LV remodelling and diastolic function in patients with T2DM. The key finding of this study was that the presence of LVH and diastolic dysfunction was common at baseline assessment in patients with T2DM and was greater in women compared with men. This sex-specific difference remained significant after a follow-up period of 25 months. The current findings also demonstrate that in patients with T2DM, pre-clinical diastolic dysfunction is associated with a higher risk of adverse outcome in women than men.

The presence of LVH and diastolic dysfunction is the hallmark of diabetic cardiomyopathy, independent of underlying coronary artery disease and hypertension. In the present cohort, nearly two-thirds of the population had adverse LV remodelling and over 20% had concentric hypertrophy. In addition, previous studies have shown a high prevalence of diastolic dysfunction in patients with T2DM, ranging from 23% to 75%.^5,16–18 Differences in patient demographics and the definition of diastolic dysfunction may account for this broad variation. Classification of diastolic dysfunction according to the latest recommendations¹⁵ has been shown to be more accurate and provide better prognostic stratification than the previous 2009 recommendations¹⁹; nonetheless, the prevalence of diastolic dysfunction in patients with T2DM, classified according to the latest recommendations, has not been evaluated. This study is the first to evaluate patients with T2DM according to the current recommendations and demonstrated that about one-quarter had diastolic dysfunction. Importantly, this study confirmed that women with T2DM had a higher prevalence of LVH and diastolic dysfunction than men. A prior study from the Framingham Heart Study has suggested that diabetes have a significant association with LV remodelling in women but not in men.⁹ This finding is consistent with another study showing that being women was an independent predictor of LVH and diastolic dysfunction in patients with T2DM.¹⁷ Further, in a recent study that provided cluster analysis of echocardiography, women with T2DM are more likely to be associated with LV diastolic dysfunction, even in the absence of LVH.¹⁰ One possible explanation for the susceptibility for women to be associated with impaired myocardial structure and function is their higher risk of comorbidities, such as higher age and obesity^20,21 than men. Further, the sex-specific gender difference in LV remodelling could be explained by the lower body size of women consist with a smaller heart, which results in a greater LV sensibility to the adverse effects of myocardial fibrosis.²² Sex differences in myocardial substrate metabolism may also play a role as women tends to have a lower glucose uptake and utilization coupled with a greater fatty acid uptake and metabolic inefficiency compared with men.²³ In this study, no significant demographic differences were noted between both sexes; this may suggest that female gender itself, rather than relative comorbidities, contributed to LVH and diastolic dysfunction in T2DM. The reason for the observed higher prevalence of LVH and diastolic dysfunction particularly in women with T2DM compared with men are likely to be multifactorial that merits further evaluation.

The assessment of PP can be simply measured and has been shown to be a surrogate of pulse wave velocity to evaluate arterial stiffness.²⁴ In this study, PP was higher in women than men and is consistent with a previous study showing that women have a higher baseline PP and average annual rate of PP change than men in both African-Americans and Whites.²⁵ The increase in systolic blood pressure and reduce in diastolic blood pressure, reflects by a higher PP, will lead to a reduction of coronary perfusion, myocardial ischaemia, and deterioration of diastolic function.²⁶ Increase arterial stiffness will further elevate cardiac afterload, which contributes to higher prevalence of LVH in diabetic women.²⁶ In addition, PP can be considered as an assessment for ventricular-arterial coupling that can reflect diastolic function and provides prognostic value to refine risk stratification in patients with cardiovascular diseases.²⁶ Due to the small study population, we were unable to evaluate the role of PP changes in relation with the alteration of LV diastolic function and MACE. Future prospective studies with a larger study population are needed to explore the possible impact of the sex-specific difference of PP changes on longitudinal changes of LV remodelling and diastolic dysfunction.

One of the novelties of this study was the availability of data on longitudinal changes of LV remodelling and diastolic function in patients with T2DM. The entire cohort had a significant increase in LV geometry including IVSd, LVPWd, LVMi, and a higher prevalence of concentric hypertrophy. In a study that involved patients with HFpEF (38% diabetic), female gender was associated with abnormal LV geometry at baseline but was similar for both sexes at 36-week follow-up.²⁷ This finding may suggest that the myocardial structural and functional differences between women and men attenuate with time. In contrast, this study involving only patients with T2DM demonstrated that women continued to have a higher prevalence of LVH and diastolic dysfunction at follow-up. This suggests that in the presence of T2DM, sex differences in myocardial structural alteration, and dysfunction are sustained over time. The possibility of greater worsening of cardiac structure and diastolic function in women compared with men over time nonetheless requires validation by studies with longer follow-up.

Women with heart failure generally have better survival than men, in particular, those without T2DM in whom female gender is protective.^28,29 However, women with T2DM has been shown to have a higher cardiovascular mortality in the Framingham Heart study.³⁰ Studies have further replicated this finding and demonstrated that T2DM attenuates this sex-specific protection in the presence of heart failure.^8,31 Extending this finding, this study demonstrated that the risk of adverse events is similarly high for both women and men with T2DM. Of interest, sex differences in terms of the prognostic value of diastolic function were noted; women had a greater than six-fold increased risk whereas men had a two-fold increased risk of adverse events when diastolic dysfunction was present. Indeed, the present finding is similar to that of the Framingham Heart Study that showed frequency of heart failure to be five times greater in diabetic women but only twice as high in diabetic men compared with age-matched control subjects.³⁰ Our finding concurs with a previous study on T2DM patients that demonstrated a similarly high cardiovascular events rate in women with diastolic dysfunction, even in the absence of LVH, compared with men with LVH and systolic dysfunction.¹⁰ Based on results from previous studies and the present observations, the increased risk of adverse events in women compared with men could be partly explained by the higher risk of LVH and diastolic dysfunction. Further, previous reports have shown that women tend to receive less evidence-based treatment than men.^31,32 Intensive clinical surveillance to detect myocardial function in order to provide appropriate intervention should be considered for women with diastolic dysfunction.

Clinical implications

Cardiovascular disease continues to be the leading cause of death in women.³³ In particular, women are twice more likely than men to develop HFpEF, characterized by the presence of LVH and diastolic dysfunction.³⁴ The current study confirmed the high prevalence of LVH and diastolic dysfunction in patients with T2DM, more pronounced in women than men. Further, a significant worsening of LVH and diastolic dysfunction occurred at follow-up in both sexes. Importantly, the risk of adverse events was high in women once diastolic dysfunction had developed. This suggests that regular echocardiographic screening is required to detect subclinical cardiac structural abnormalities and diastolic dysfunction to improve risk stratification in T2DM. Although the current guideline does not advocate routine echocardiography screening in otherwise asymptomatic T2DM patients,³⁵ we believe that early detection of diastolic dysfunction by non-invasive methods is essential to improve risk stratification in these patients. The use of novel echocardiography techniques, such as speckle tracking derived myocardial strain analysis,^36,37 may further improve risk stratification. Large studies are warranted to identify the most health economical and accurate method to detect women with T2DM at risk of adverse events.

Limitations

Although this study is by far one of the largest to examine longitudinal changes to cardiac structure and function in patients with T2DM, the sample size may still have been underpowered to enable evaluation of differences between sexes, especially the possible greater worsening of LV remodelling and diastolic dysfunction in women compared with men at follow-up. Further, advanced strain analysis for both systolic and diastolic function was not performed and requires validation by future studies. In this study, survival analysis for prediction of MACE was based only on the clinical and echocardiographic baseline assessment. Future prospective studies are warranted to evaluate the sex-specific prognostic value of the changes of diastolic dysfunction, PP, and HbA1c in patients with T2DM. Prior studies have indicated that microvascular abnormalities could influence the cardiac structure and function for patients with T2DM, and increased the risk of cardiovascular events,^38–40 However, this study did not systematically evaluate patients’ microvascular complication. Future prospective study is thus required to explore the influence of proteinuria/microalbuminuria, neuropathy, and retinopathy on MACE and longitudinal changes of cardiac structure and function in patients with T2DM. The current study population comprised only patients with T2DM who were asymptomatic; future studies that include patients with established cardiovascular complications will need to be conducted to confirm our findings.

Conclusion

This study demonstrated prominent sex differences in the pattern of LVH and diastolic dysfunction in patients with T2DM. In particular, the prevalence of LVH and impaired diastolic function was greater in women than in men. The sex difference in cardiac structural alteration and diastolic dysfunction remained significant at follow-up. Importantly, pre-clinical diastolic dysfunction in these patients was independently associated with MACE in women but was neutral in men.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Funding

This work was partially supported by the Summit Grant from the Guangdong and Shenzhen Ministry of Health, Grant Number: SZSM201911020.

Conflict of interest: none declared.

References

Author notes