Background
Obesity and T2D are well-accepted risk factors for the development of left ventricular hypertrophy (LVH) [
1,
2]. However, it remains unclear as to whether these stresses are sufficient to cause LVH, independent of hypertension and other cardiac disease. This is largely due to the asymptomatic nature of LVH and the way in which risk is monitored in normotensive obese and/or T2D patients.
LVH is traditionally characterised by increased thickness of the LV posterior wall diameter (LVPWd) and/or increased LV mass [
3]. There are three LV geometric patterns that can identify the type of LVH present, eccentric hypertrophy (increased LV mass), concentric remodelling [increased relative wall thickness (RWT) of the LVPWd, normal LV mass] and concentric hypertrophy (increased LV mass and increased RWT). These changes in cardiac structure are often accompanied by diastolic dysfunction (DD, impaired LV relaxation) and can be detected in obese patients via transthoracic echocardiography (TTE) [
3]. However, requests for patients to undergo TTE are often restricted to those with chronic hypertension, or to those who have returned an abnormal electrocardiogram result after presenting with symptoms such as arrhythmia, shortness of breath or chest pain. Monitoring risk in obese and/or T2D patients this way presumes that without hypertension, obesity and T2D are insufficient stresses to induce LVH.
Major limitations of past studies are that obesity and T2D have often been considered one disease state, and the individual contributions of these diseases to LVH in the absence of hypertension have remained obscured. Furthermore, in studies that have appropriately distinguished obesity [
4,
5] and T2D [
6‐
8], there has been a propensity to use LV mass alone as an independent variable to detect LVH, without assessing the RWT of the LV posterior wall. This restricts the type of LVH that can be detected to eccentric hypertrophy and prevents the detection of concentric remodelling and concentric hypertrophy. Previous studies suggest there is prognostic value in assessment of LV geometric patterns, with increased LV mass and concentric hypertrophy in particular [
9,
10] found to associate with increased risk of adverse cardiovascular events. Whether this prognostic value applies to normotensive patients with obesity and T2D remains unresolved.
The current study aimed to determine whether patients presenting with obesity and type 2 diabetes, either alone or in combination, exhibit LVH in the absence of hypertension. Both traditional indices of LVH and LV geometric patterns were used to identify LVH. Furthermore, we sought to determine whether LV geometry predicted the presence of diastolic dysfunction and whether routine metabolic parameters can be used to predict LVH in these patients.
Methods
Study approval
This cross sectional study comprised a total of 353 Caucasian patients, from the University Hospital Geelong and the Geelong Endocrinology and Diabetes Centre. The study was conducted in accordance with National Health and Medical Research Council (NHMRC) guidelines and was approved by the Human Research Ethics Committee (HREC) via the Barwon Health Research and Integrity Unit, in accordance to guidelines outlined in section 5 of the National Statement on Ethical Conduct in Human Research.
Participant groups
Based on the diagnosis of obesity, T2D and hypertension, patients were designated into one of the following groups; normotensive obese (N.Obese) n = 58, normotensive T2D (N.T2D) n = 41, normotensive obese/T2D (N.Obese/T2D) n = 42, hypertensive obese (H.Obese) n = 71, hypertensive T2D (H.T2D) n = 74 and hypertensive obese/T2D (H.Obese/T2D) n = 67. For the diagnosis of obesity, T2D and hypertension, basic clinical and metabolic data were collected for each participant consisting of age, sex, height, weight, blood pressure, HbA1c %, fasting glucose, LDL-C, HLD-C, cholesterol and triglyceride levels (after overnight, 8 h minimum fast), any history of anti-hyperglycaemic or anti-hypertensive medication, and any history of cardiovascular and/or systemic disease.
Characterisation of and inclusion/exclusion criteria for participant groups
Obesity was characterised as a BMI ≥30 kg/m
2. Patients with a history of anti-hyperglycaemic medication were permitted within the obese group, due to the increasingly common use of biguanides in pre-diabetic/obese patients. T2D was characterised as having three or more elevated fasting glucose levels within a 12-month period, of ≥7 mmol/l, with or without a history of anti-hyperglycaemic medication. Hypertension was characterised as having both elevated diastolic and systolic blood pressure of ≥140/90 mmHg, with or without a history of anti-hypertensive medication. Those patients with controlled hypertension (i.e. history of hypertension or use of anti-hypertensive medication, with a blood pressure <140/90) were excluded from the study. Additional exclusion criteria for patients of all groups included history of cardiac disease or systemic disease, age <18 years, or the collection of accompanying clinical and metabolic medical records within a time period >6 months before or after TTE procedure. For a more detailed description of the study design see Additional file
1.
Blood pressure measurements
Blood pressure (BP) was recorded as per the American Heart Association guidelines, with the use of OMRON Intelli sense HEM-907 or HBF-1300 and cuff bladder at least 80% of the patient’s arm circumference. In the incidence of an elevated BP reading (≥140/90 mmHg), the measurement was repeated up to three times. With the lowest BP measurement recorded. Pulse pressure mmHg was calculated by subtracting diastolic BP from systolic BP (systolic BP mmHg–diastolic BP mmHg).
To separate normotensive obese patients based on metabolic health. We adhered to Karelis criteria. With metabolically healthy patients determined as; fasting glucose ≤5.5 mmol/l, HDL-C ≥1.4 mmol/l, LDL-C ≤2.6 mmol/l, cholesterol ≤5.5 mmol/l and triglycerides ≤1.8 mmol/l. Patients were categorised as being metabolically unhealthy if they exhibited >1 more parameter outside these normal ranges.
Transthoracic echocardiography
Sonographers were qualified with a Diploma of Medial Ultrasonography or equivalent. Both the sonographers that performed the echocardiography and cardiologists that analysed the results were blinded to the study groups, due to the retrospective nature of the study. All echocardiograms were performed using the Phillips Ie33 with a S5-1 transducer. A combination of two dimensional, M-mode, pulsed wave and continuous wave Doppler and tissue Doppler were used. Left ventricular diameter and wall thicknesses were measured in the parasternal long axis view using two-dimensional or M-mode measurements [left ventricular internal diastolic dimension (LVIDd), left ventricular internal systolic dimension (LVISd), interventricular septum dimension (IVSd), left ventricular posterior wall dimension (LVPWd)]. Of note, while M-mode was used to measure the LV wall thickness whenever possible, in cases where the M-mode was not able to be properly aligned (orthogonal) two dimensional echocardiography was used. Mitral inflow velocities (E’ velocity, Peak E-wave, Peak A-Wave) and deceleration times (DT) were measured using pulsed wave Doppler in the apical 4 chamber view. Echocardiographic data was analysed using proprietary software.
Characterisation of diastolic dysfunction
Diastolic dysfunction (DD) was characterised according to the American Society of Echocardiography (ASE) guidelines [
11]. Patients were graded with either normal diastolic function (E′ ≥ 10 cm/s) or DD, characterised as Grade 1 (impaired relaxation) E′ < 10 cm/s, E/A < 0.8, E/E′ ≤ 8; Grade 2 (pseudonormal) E′ < 10 cm/s, E/A 0.8–1.5, E/E′ 9–14; or Grade 3 (restrictive) E′ < 10 cm/s, E/A ≥ 2, E/E′ > 14.
Left ventricular geometry
LV mass was estimated according to ASE guidelines [
12], in which LV mass (grams) = (0.8·[1.04·(LVEDd + IVSd + LVPWd)
3 − (LVEDd)
3]) + 0.6). LV mass was then indexed to body surface area (BSA, g/m
2) and to height (g/m
2.7). RWT was calculated using the formula, RWT = ((IVSd + LVPWd)/LVEDd) and via ((2·LVPWd)/LVEDd). LV geometry was characterised using the following criteria; Normal LV geometry, RWT ≤ 42, LVMI (g/m
2.7) ≤51; eccentric hypertrophy (EH), RWT ≤ 42, LVMI (g/m
2.7) >51; concentric remodelling (CR), RWT > 42, LVMI (g/m
2.7) ≤51 and concentric hypertrophy (CH), RWT > 42, LVMI (g/m
2.7) >51.
Statistical analysis
Continuous variables were represented as means ±1 standard deviation (SD), unless otherwise stated. Means of continuous variables were analyzed via ANOVA assessed with Bonferroni, and associations were determined by performing linear regression analysis, assessed with Pearson’s correlation coefficient. Categorical variables were expressed as percentages or prevalence and analyzed via Chi square tests, using fisher’s exact test. To determine independent predictors of categorical variables, multivariable logistic regression analysis was performed with variables adjusted for as detailed. Of note, aortic stenosis was not adjusted for as the presence of aortic stenosis (characterised by an ascending aorta of <3.7 cm) did not associate with diastolic dysfunction or LV geometric patterns in the normotensive patients. p < 0.05 was considered significant. All statistical analysis were performed using SPSS version 23.
Discussion
Despite professional knowledge of the risks associated with obesity and T2D in the development of cardiovascular disease, monitoring blood pressure remains the first line method when assessing risk in normotensive obese and/or T2D patients. While this practice allows for the early detection of hypertension, it may not be sufficient to allow for the early detection of LVH and DD. Indeed, results from this current study suggest that in the absence of hypertension, LVH and DD may be present and detectable via TTE in normotensive obese and/or T2D patients. The major findings from this study are that; (1) Significant alterations in LV remodelling indicative of LVH were detected in normotensive metabolically non-healthy obese, T2D and obese/T2D patients. (2) Assessment of LVPWd via recommended ASE guidelines underestimated the presence of LV remodelling. (3) Concentric remodelling was a predictor of grade 1 DD and concentric hypertrophy and RWT predictors of grade 3 DD, independent of sex, age and BMI. (4) Normotensive patients with an increased risk of a RWT > 42 with those exhibiting pulse pressure ≥54 mmHg.
Considering obese patients first, metabolically healthy N.Obese patients exhibited a relatively low prevalence of LV remodelling, with >80% of these patients exhibiting normal LV geometry. It was only in those N.Obese patients characterised with “poor” metabolic health that significant alterations in LV geometry were detected. This was particularly true for concentric remodelling (47%), which was 9.4 times higher than that in metabolically healthy N.Obese patients and 1.6 times higher than that in H.Obese patients. This suggests that obesity per se, as defined by a BMI ≥ 30 (kg/m2), does not promote the development of concentric remodelling, but rather it is the metabolic health status of the patient with a BMI ≥ 30 (kg/m2) that has the greatest influence on increases in RWT. Supporting this claim, our study showed that BMI had no association with RWT or diagnosis of concentric remodelling. Therefore, these results suggest additional risk assessment should be performed in normotensive obese patients with more than one of the following metabolic parameters; fasting glucose >5.5 mmol/l, HDL-C <1.4 mmol/l, LDL-C >2.6 mmol/l, cholesterol >5.5 mmol/l and triglycerides >1.8 mmol.
H.Obese patients on the other hand exhibited not only concentric remodelling (28%) but also concentric hypertrophy (25%), at a prevalence 4.8 and 2 times higher than that in metabolically healthy and non-healthy N.Obese patients respectively. This is consistent with previous studies in which autopsy results identified hypertensive obese patients to exhibit both concentric remodelling and concentric hypertrophy and for normotensive obese patients to present with mainly concentric remodelling [
17]. N.T2D patients exhibited similar alterations in LV geometry as metabolically non-healthy obese, showing a comparable prevalence of concentric remodelling (44%) and a low prevalence of concentric hypertrophy (4%), 6.75 times lower than that in H.T2D patients. Interestingly, it was only in those N.Obese/T2D patients in which the predominant LV geometric patterns detected included both concentric remodelling (30%) and concentric hypertrophy (36%), which was comparable to that in H.Obese patients. These results suggests that the co-existence of obesity and T2D in normotensive patients had an effect on the development of CH which was similar to the negative influence hypertension had when coupled with obesity or T2D alone.
To our knowledge, this is the first study reporting that the use of unadjusted LVPWd vs RWT underestimated the presence of LV remodelling in obese and T2D patients. This had particular relevance in normotensive patients, with 45% of those characterised with concentric remodelling exhibiting LVPWd within the ranges of normal ASE guidelines. This result was unchanged regardless of how RWT was derived. However, consideration of septal wall thickening was preferred due to the presence of asymmetric LV remodelling in some patients, with IVSd showing similar patterns of enlargement as LVPWd within our groups (Additional file
1: Table S1). Of note, this increase in septal thickening was accompanied by aortic stenosis, with a thinning of the proximal ascending aorta correlating with increased IVSd (r = 0.24, p = 0.001, data not shown), a correlation that has previously been identified in subjects with asymmetric remodelling [
18]. These data suggest that in normotensive obese and/or T2D patients assessment of unadjusted LVPWd alone may underestimate the presence of LV remodelling (both symmetric and asymmetric) and, as such, the additional use of RWT may provide a more sensitive measure in these patients.
The implications of underestimating LV remodelling are evident in this study with the use of multivariable logistic regression analysis identifying an association between concentric remodelling with diagnosis of grade 1 DD and concentric hypertrophy and RWT with grade 3 DD (independent of sex, age and BMI). This suggests that reliance on unadjusted LVPWd in normotensive obese and T2D patients underestimates not only LV remodelling, but also associated cardiovascular risk. While previous studies have identified associations between LV geometry with adverse cardiac function and cardiovascular risk [
19‐
24], these studies have predominately focused on hypertensive patients or have not had access to subject data regarding hypertension. The data from this study provides insight into the prognostic value of LV geometry, specifically in normotensive obese and T2D patients.
While subject numbers in the present study were low, the identification that pulse pressure and fasting glucose to be associated with RWT may assist health practitioners in narrowing down the otherwise large patient pool of at risk individuals to which this study relates, with pulse pressure ≥54 mmHg and fasting glucose ≥7.7 mmol predictors of RWT > 42. RWT was chosen as an independent variable to detect risk of LVH due to concentric remodelling and concentric hypertrophy being the predominant LV geometric patterns detected for which both have the common requirement for increased RWT. In addition, RWT provided a more sensitive measure to detected LV remodelling in our patients and increases in RWT have previously been associated with adverse cardiovascular events [
9,
25,
26]. In healthy subjects normal pulse pressure levels have been detected at 40 mmHg. The recorded pulse pressure measurements in the current study were increased, averaging >50 mmHg in normotensive and >60 mmHg in hypertensive patients. Similar increases in pulse pressure have previously been recorded in subjects with metabolic syndrome [
27,
28]. To our knowledge, this is the first study identifying the use of pulse pressure as a predictor of increased RWT that can be applied to all normotensive obese, T2D and obese/T2D patients. In a dataset from the HyperGEN study [
29] normotensive subjects with a higher pulse pressure (>60 mmHg) were associated with an increased prevalence of thickening of the LVPWd and increased RWT. This pulse pressure quartile however contained a lower average BMI (27.06 kg/m
2) and percentage of T2D subjects (3.97%) compared to current study.
Fasting glucose, has been associated with indices of LVH in the past [
30]. Furthermore, patients with impaired fasting glucose or T2D have previously been reported to exhibit a 9 and 11% respectively increase in the prevalence in concentric remodelling than that in euglycemic patients (of note, 46% of patients were hypertensive) [
31]. The identification of the association with fasting glucose with RWT in the current study was strengthened when accounting for age and BMI. When comparing the use of fasting glucose vs pulse pressure as a predictor of increased RWT via multivariable stepwise regression analysis, fasting glucose was a stronger predictor of RWT (data not shown). In regards to the practicality of using pulse pressure and fasting glucose as markers of increased risk, these parameters are already routinely collected in normotensive obese and T2D patients, therefore their use will not result in an additional burden in either cost or time. However, it should be noted that although the presence of elevated pulse pressure and fasting glucose were associated with LV remodelling, this study is not suggesting that these are a requirement and, as such should not provide a means to exclude patients at risk.
There are some limitations that need to be considered when interpreting the findings from this study. In regards to diastolic function, the reader is reminded that these measurements were obtained from an assessment at one time point and that diastolic function may vary between examinations. Furthermore, the size of the research groups were limited, more hypertensive vs normotensive patients were included in the study and analysis was restricted to patients of one ethnic group. The strengths of the study included providing clearly separated normotensive and hypertensive obese, T2D and obese/T2D patients. Allowing us to tease out the influences of these conditions individually and in combination. Furthermore, despite the restricted patient numbers in the study we ensured that our dataset exhibited the normal associations between age with diastolic decline and LV structure, adding confidence that the findings from this study are relevant to the wider population. In addition, we have focused our analysis on echocardiography techniques and clinical measurements that are already routinely measured in day to day practice, increasing the application potential of the study.
Authors’ contributions
Study concept and design; KADJ, AA, KC, MK, and SLM. Data collection; AA, KC and MK. Data acquisition; KADJ, AA, KC, MK, and SLM. Reformatting acquired data for analysis; KADJ, SS, KM. Data analysis; KADJ, JKC, GDL, AA, KC, MK and SLM. Wrote the manuscript; KADJ and SLM. Critical analysis of manuscript; KADJ, JKC, GDL, AA, KC, MK and SLM. All authors approved final manuscript. All authors read and approved the final manuscript.