Heart failure is only one of several possible underlying causes of the broader issue of circulatory failure in  which an abnormality of some component of circulation is responsible for inaequate cardiac output.  Circulatory failure can arise from heart failure, or from insufficient blood volume, low concentrations of oxygenated hemoglobin in arterial blood, or other circumstance that prevent an otherwise healthy heart from adequately meeting the metabolic needs of the body.

The assessment of left ventricular hypertrophy in hypertension

Abstract

The presence of left ventricular hypertrophy (LVH) in hypertension, as detected by the electrocardiogram or echocardiography, is associated with an increased risk of mortality and morbidity several times above and beyond the risk of hypertension alone. The LIFE (Losartan Intervention For Endpoint reduction in hypertension) study confirmed that pharmacological agents, which reduce LVH, confer further reduction in morbidity and mortality. This makes the identification of patients with LVH all the more important. In this article we describe the various methods available to diagnose the presence of LVH in patients with hypertension, and consider their strengths and their place in clinical practice and in research.

Introduction

Hypertension is associated with an increased risk of morbidity and mortality as a result of an increase in the incidence of myocardial infarcts, strokes, cardiac and renal failure, as well as sudden death . The presence of evidence of left ventricular hypertrophy (LVH) on the electrocardiogram (ECG) increases the risk of mortality and morbidity several times above and beyond the risk of the hypertension alone . Echocardiography provides a more sensitive test for identifying LVH and elevated left ventricular mass (LVM), as measured by M-mode echocardiography, which is also associated with increased cardiovascular risk . The two tests identify different populations of patients with LVH and the combination of the two tests provides additive prognostic information   The ECG criteria of LVH identify a group of hypertensive patients with a particularly high LVM . This is clinically relevant as there is an incremental relationship between LVM index and cardiovascular risk. Interestingly, the continuous relationship between LVM index and cardiovascular risk is present even in the quintiles below the upper limit of normal for LVM index, as measured by echocardiography (125 g/m² for men and 110 g/m² for women) .There is recent evidence that regression of LVM, through pharmacological intervention, reduces the cardiovascular risk   Furthermore, the LIFE (Losartan Intervention For Endpoint reduction in hypertension) study confirmed that pharmacological agents that reduce LVH above and beyond lowering of blood pressure (BP), confer further reduction in morbidity and mortality It is therefore important to identify patients with LVH, both for prognosis and for tighter BP control. Currently, the European Society of Hypertension guidelines recommend that an ECG should be performed to assess for the presence or absence of LVH based on the Sokolow–Lyons or the Cornell Product criteria. They recommend echocardiography, using M-mode to measure wall thickness and calculate LVM index, when treatment decisions are uncertain, to refine the classification of the overall risk

This article describes the various methods available to diagnose the presence of LVH in patients with hypertension, both in the context of clinical practice and research. The optimal method should be easy to perform, inexpensive, widely available and accurate. We consider how the various methods compare.

The electrocardiogram

The ECG is easy to perform, widely available and inexpensive. It is usually the first test to detect LVH, on the basis of numerous validated criteria of variable complexity. The sensitivity of these criteria varies in published studies depending on the severity of hypertension and hence prevalence of LVH in the cohort being studied   However, these criteria generally have high specificities but low sensitivities . In the validation studies of these criteria, the sensitivity has been reported to be as high as 50% in severely diseased necropsy populations, to as low as 6–17% in Cornell and Framingham studies . Some of these criteria are simple to use, like the Sokolow–Lyon voltage criterion which is based on voltage amplitude of leads SV₁ + RV₅ or RV₆ , or the Lewis index voltage criterion which is based on voltage amplitudes in leads I and III [(RI – RIII) + (SIII - SI) >= 17 mm]  The Cornell voltage criterion is another simple, gender-specific criterion, and is based on the voltage amplitude sum of R in lead aVL plus S in lead V₃ being >= 28 mm for men and >= 20 mm for women   Other criteria are more complex and are based on scoring several ECG findings, such as the Romhilt–Estes score, which is based on a point system where six different criteria are awarded a different number of points with a maximum of 13 points and two cut-off values of 4 and 5 points, with the 5-point cut-off value being more and specific [ The Romhilt–Estes score was validated with very high sensitivity and specificity in 90 severely hypertrophied hearts from post-mortems   The Perugia score is another composite criterion that requires positivity in one of three different criteria: Cornell with a cut-off >= 24 mm for men; Romhilt–Estes score with a cut-off value of 5 points; and left ventricular strain pattern . In a study by Verdecchia et al. the Perugia score was associated with the highest rate of cardiovascular events and mortality when compared with Cornell voltage, Romhilt–Estes score, left ventricular strain and Sokolow–Lyon criteria in a 10-year follow-up.In the LIFE study, two ECG LVH criteria were used to confirm regression of LVH: Cornell product criterion and Sokolow–Lyon voltage criterion . The Cornell product criterion is based on the product of the Cornell voltage, with the addition of a 8 mm adjustment for females, and the QRS duration [(RaVL + SV₃) × QRS duration] > 2440 mm/ms . Based on the LIFE study, the European Society of Hypertension guidelines recommend the use of the Sokolow–Lyon voltage and the Cornell product criteria for the assessment of LVH. However, the LIFE study investigators selected a partition value for the Sokolow–Lyon voltage criterion of 38 mm for both men and women. It is recognized that there are gender differences in QRS duration and voltage measurements, which are not completely accounted for by differences in body size and LVM, and hence gender-specific partition values are important . Recent data suggest that the partition value for the Sokolow–Lyon voltage criterion should be 34 mm for females and 38 mm for males   This postulates that the use of a single partition value of 38 mm in the LIFE study may have resulted in a lower sensitivity at detecting LVH in females.

A Sokolow–Lyon product criterion, based on the product of the Sokolow–Lyon voltage and the QRS duration [(SV₁ + RV₅ or RV₆) × QRS duration], has been described , with gender-specific partition values of 4000 mm/ms in males and 3000 mm/ms in females suggested recently  Both Sokolow–Lyon product and Cornell product criteria are thought to be more sensitive at detecting LVH than their respective voltage criteria

The ECG criteria, however, have significant inter-study variability, which makes them less reliable for temporal follow-up in individual patients . The large numbers recruited into studies such as LIFE overcome this problem.

Echocardiography

Echocardiography is also relatively easy to perform and widely available. It is more expensive than the ECG but provides additional information on the structure and function of the heart and valves. Both M-mode and two-dimensional (2-D) echocardiography are used in the measurement of LVM. M-mode has the advantage of superior endocardial definition, as the high frame rate improves the resolution. It is the most widely used echocardiography method to measure LVM because it was the first to be validated and because it is relatively simple and quick. The M-mode method measures the left ventricle (LV) in one dimension and assumes a prolate ellipsoid shape for the LV with a ratio of long to short axis lengths of 2: 1 The calculation of LVM is based on a mathematical formula, LVM = 0.8(1.04[(LVIDD + PWTD + IVSTD)³ - (LVIDD)³]) + 0.6 g, as modified by Devereux et al. using the American Society of Echocardiography (ASE) convention, where LVIDD is left ventricle internal dimension in diastole, PWTD is posterior wall thickness in diastole and IVSTD is intraventricular septal thickness in diastole. This modified formula was validated on necropsy findings in 52 subjects

However, the accuracy and reproducibility of M-mode at measuring LVM has been debated. This is a result of the potential for variations in measured wall thickness, depending on the ultrasound beam angle to the LV wall and the assumption that the wall thickness is uniform throughout the LV. In addition, the assumed prolate ellipsoid shape of the LV is no longer valid in patients with LVH . Furthermore, LVM as measured by M-mode echocardiography relies on linear measurements of wall thickness which, when cubed, increases the standard deviation (SD) by a factor of 2–3. As elevated LVM is defined as mean + 2SD, M-mode echocardiography will underestimate the prevalence of LVH in hypertensive cohorts Finally, the cubing of the measurements for LV wall thickness amplifies any errors of such measurements, resulting in a significant variation of LVM estimates compared to direct measurement by three-dimensional (3-D) cardiac magnetic resonance imaging (MRI)

Two-dimensional echocardiography is thought to be more accurate and reproducible than the M-mode method .Two-dimensional echocardiography takes into account the length of the LV as well as the myocardial wall thickness, but is limited by the lower frame rate and hence lower resolution. It is important to ensure that the LV apical view is not foreshortened, as it would change the length of the LV and hence the LVM. Two-dimensional echocardiography relies on mathematical formulae to estimate the LVM but there is no cube function in these formulae and hence measurement errors are not magnified. However, it still assumes a prolate ellipsoid shape of the LV and a uniform LV wall thickness. Two-dimensional echocardiography is less widely used to estimate the LVM than M-mode echocardiography, as it is more time consuming and it is more difficult to obtain images of suitable quality.

Tissue harmonic imaging (THI) is a relatively new ultrasound imaging modality with an improved resolution in comparison to fundamental imaging, which is the standard ultrasound imaging modality. THI improves the image quality and the detection of the endocardial border, and has been shown to be superior at measuring ejection fractions compared to fundamental imaging .Because of the improved image quality, THI is becoming the default imaging modality in some echocardiography units. However, THI overestimates M-mode echocardiography measurements of LV wall thickness as well as M-mode LVM .Normal ranges for echocardiography should be updated to take account of the shift to THI.

Indexed left ventricular mass

There are a few indexes to correct for body size. These are height, weight and body surface area (BSA). These different indexes result in different prevalence of LVH in the population under investigation .Indexation to BSA reduces variability due to body size and gender, and is the most widely used method in the literature .but it can underestimate LVM at the upper end of the distribution .Indexation of LVM to height is thought to be more accurate in the obese. Regression models for height^2.7 provide the most accurate estimation for LVH and cardiovascular risk .The association between obesity and hypertension makes this relevant.
Geometric patterns

Different geometric forms of LVH have been adopted to classify the maladaptive responses of the LV in hypertension . LV geometry is classified into four exclusive groups on the basis of LVM and relative wall thickness: concentric hypertrophy (increased mass and increased relative wall thickness), eccentric hypertrophy (increased mass and normal relative wall thickness), concentric remodelling (normal mass and increased relative wall thickness) and normal geometry (normal mass and normal relative wall thickness). Krumholz et al. investigated 3200 patients from the Framingham study, indexing LVM to height, and found that patients with concentric hypertrophy (defined by thickness of the septum or the posterior wall divided by the LV radius at end-diastole >= 0.45) have a higher rate of cardiovascular events and mortality compared to patients with eccentric hypertrophy. Further studies by Verdecchia et al. found a higher risk, beyond LVM indexed to BSA, for concentric remodelling, but no additional risk for those already identified as having concentric hypertrophy. This implies that simply measuring the LV wall thickness and LV radius offers the same, if not possibly more, prognostic information as measuring LVM.

Three-dimensional echocardiography

Three-dimensional echocardiography had been limited by time-consuming sequential acquisition of multiple 2-D image planes using transoesophageal or transthoracic approaches, as well as time-consuming post-processing to reconstruct 3-D datasets from multiple 2-D images. While 3-D echocardiography improved the intra- and interobserver variability of the LVM measurements compared to M-mode and 2-D echocardiography , with its accuracy reported to be close to that of cardiac MRI   the technique remained limited to a few research centres. However, recently 3-D echocardiography has advanced significantly, with the introduction of second-generation real-time 3-D technology. The second-generation matrix array probes, with 3000 simultaneously active ultrasound elements, have solved the acquisition difficulties, while advances in computer technology and analysis software have shortened the duration of post-processing  . Quantification of LVM using real-time 3-D echocardiography was found to be highly reproducible and to have excellent correlation with cardiac MRI  In contrast to M-mode and 2-D echocardiography, 3-D echocardiography does not rely on geometrical assumptions for calculating LVM   One remaining limitation for 3-D echocardiography is the acoustic windows in some subjects.

Cardiac magnetic resonance imaging

At the present time, cardiac MRI is relatively expensive, not widely available and requires more expertise than echocardiography. However, it has been shown to be the most accurate and reproducible tool for the estimation of LVM  This has been validated in animal studies  and was demonstrated to be more accurate and reproducible than M-mode and 2-D echocardiography measurements . Cardiac MRI provides a spatially defined 3-D dataset at multiple levels throughout the heart, and hence the measurement of the LVM does not require geometric assumptions about the shape of the LV. The excellent contrast between blood and myocardial tissue and the high spatial resolution mean that the endocardial and epicardial contours are easily defined ( . Currently two pulse sequences are in common use: the segmented k-space turbo gradient echo (TGE) technique and the steady-state free precession technique (SSFP). LVM measurements using these two pulse sequences have been compared and a systematic difference was found between them, with the SSFP pulse sequence yielding smaller LVM measurements than TGE   Hence, it is important to use a normal range that corresponds to the MRI pulse sequence in use

Low standard deviation left ventricular mass

The sensitivity of the ECG criteria for detecting LVH depends on the severity of hypertension and hence the prevalence of LVH in the cohort being studied   The prevalence of elevated LVM is also dependent on the severity of hypertension in the population being studied. However, it is also dependent on the methodology of calculating the LVM. The M-mode echocardiography method has a high SD, which is the result of the ‘cube function’ in the mathematical formula, which magnifies the SD . For example, the SD of the mean LVM index calculated using M-mode echocardiography for a cohort of normal individuals was 21 g/m² for men and 19 g/m² for women   The equivalent SD for cardiac MRI using TGE pulse sequence, where there is no such ‘cube function’, was 9.1 g/m² for men and 7.7 g/m² for women . The larger SD of LVM index by M-mode echocardiography results in a higher upper limit of normal (based on the upper 95th percentile of the distribution; mean + 2SD)  . This leads to subjects with milder degrees of elevated LVM falling within the normal range of M-mode echocardiography. This may explain why Schillaci et al. found an increased cardiovascular risk even in individuals below the upper limit of normal of LVM index using M-mode echocardiography. To detect individuals with milder degrees of LVH, the LVM needs to be obtained using a method with a low SD such as 2-D echocardiography, 3-D echocardiography or 3-D cardiac MRI . An alternative ‘low SD’ method is to assess the geometric patterns of the structural remodelling of the LV by measuring wall thickness and wall thickness: internal radius ratio

Implications for research

The LIFE study, where all patients had LVH on ECG, confirmed that LVH regression results in improved outcomes, independent of BP reduction Demonstration of LVH regression using ECG criteria of LVH in the LIFE study was only possible as it was a very large cohort. Demonstration of LVM regression using M-mode or 2-D echocardiographic methods also requires relatively large cohorts of patients because of the relatively high observer and interstudy variability [59]. Precise measurement of LVM using cardiac MRI or 3-D echocardiography means that a much smaller sample size is needed to detect changes in LVM. It is calculated that the number of subjects needed for such studies using cardiac MRI is as little as a tenth of the number required to do the same study using M-mode echocardiography [59]. Cardiac MRI or 3-D echocardiography should be the techniques of choice in clinical trials investigating LVM regression where the accuracy of LVM measurements will allow for the accurate detection of small degrees of change in LVM in small cohorts of patients, particularly when recruitment of large numbers of patients is not possible.

Implications for clinical practice

In clinical practice the ECG is usually the first test performed to assess for the presence of LVH. Patients who are ECG LVH positive may not require further prognostic assessment as the ECG criteria have high specificity and they identify the subgroup of patients with the highest LVM and the highest risk . However, a normal ECG does not exclude the presence of LVH, and these patients should be investigated further. The M-mode echocardiography method remains the most widely available technique and will detect all but the mildest degrees of LVH. A precise estimate of LVM or serial follow-up in an individual patient may not be accurate in view of the observer variability and the fact that any error in measurement is cubed in the M-mode method of measuring LVM. de-Simone et al. calculate that the probability of a true change in LVM over time would be detected by the same observer only if the change in LVM is greater than 18% of the initial value. The geometric classification, which depends on simple measurements of wall thickness and LV radius using M-mode, or 2-D measurements when M-mode measurements are not possible, may suffice to give an accurate and reproducible assessment of the degree and the geometric type of LVH, with the same, if not possibly more, prognostic information as measuring LVM   with the added advantage of increased sensitivity and hence the detection of individuals with milder degrees of LVH. The use of echocardiography to investigate LVH has the additional advantage of providing a quick and accurate assessment of the valves and LV systolic and diastolic function.

Cost implications

The acquisition time of the cardiac MRI scan for LVM using the most recent techniques can be as short as 3 min and the semi-automated contour detection for LVM can be achieved within a minute. The speed and accuracy of cardiac MRI may make LVM measurement using cardiac MRI cost effective in clinical practice in the near future. However, second-generation real-time 3-D echocardiography systems are likely to become more widely available in the near future. This new technology will have the accuracy and reproducibility approaching that of cardiac MRI, while maintaining a cost advantage, as well as being more patient friendly. In clinical practice, cardiac MRI may be reserved for patients with limited acoustic windows or in subjects with deformed ventricles . In the meantime, the ECG, M-mode and 2-D echocardiography remain the most cost-effective methods that are widely available in clinical practice to date