Background
Myocardial acceleration during isovolumic contraction (IVA) has been validated as a relatively load-insensitive noninvasive index of contractility. Its feasibility, reproducibility, and variation between segments have not been studied in detail, and thus its utility in clinical practice has not been established.
Methods
We analyzed myocardial velocity loops (median frame rate 182 s −1 ) from 20 young volunteers (10 men, aged 25.7 ± 2.9 years), 20 patients with type 2 diabetes (14 men, aged 64.1 ± 8.5 years), and 20 patients with heart failure (17 men, aged 64.6 ± 7.7 years). Long-axis IVA was measured in all walls at the annulus and in basal and mid-ventricular segments. Intraobserver reproducibility for 1 observer in all subjects and interobserver reproducibility among 3 observers in 10 subjects from each group were assessed.
Results
In control subjects, subjects with diabetes, and subjects with heart failure, the feasibility of measuring IVA was 97%, 89%, and 82%, respectively; intraobserver reproducibility was 12%, 18%, and 30%, respectively (pooled coefficients of variation); and mean interobserver reproducibility was 23%, 21%, and 28%, respectively. IVA was lower in the mid-ventricular segments by 24% to 43% compared with the annulus, and IVA was higher in the right than the left ventricle ( P < .001). IVA of the medial mitral annulus discriminated those with heart failure from those with diabetes and controls, and had acceptable intraobserver reproducibility across groups (mean coefficient of variation 13%).
Conclusion
IVA may be used as a research tool if it is measured at the medial mitral annulus, but its clinical applicability is hampered by low reproducibility, especially in patients with impaired left ventricular function in whom it would otherwise be most useful.
Noninvasive diagnosis of myocardial contractility has been an elusive goal, because most indices of left ventricular (LV) function are altered by changes in loading. Isovolumic acceleration (IVA) has been proposed as an index of right ventricular (RV) and LV contractile function that is relatively insensitive to changes in loading conditions, at least within physiologic ranges. IVA correlates closely with invasively derived hemodynamic indices of contractility, such as ±dp/dt and end-systolic elastance.
The potential use of IVA has been demonstrated in research studies for the diagnosis of myocardial ischemia and myocardial dysfunction in various other diseases. Some investigators reported that IVA is dependent on loading conditions but only when these were altered beyond physiologic levels. Furthermore, IVA was conceived as a regional measurement that gives information about global LV contractile function, but regional variations have been reported from studies in animals and humans.
Data are scarce regarding the clinical application of IVA in the assessment of subclinical or overt systolic myocardial dysfunction. A thorough description of the feasibility and reproducibility of this parameter over a large spectrum of myocardial dysfunction is therefore needed. The reproducibility of IVA has been reported to be in the range of 10% to 25% (coefficients of variation [CV]), but the published studies were small. However, one recent study performed in human infants reported poor reproducibility of IVA (CV > 40%).
The objectives of this study were to assess regional variations of IVA in the LV and RV, and to study the feasibility and intra- and interobserver reproducibility of analyzing IVA over a wide spectrum of clinical circumstances, ranging from normal subjects to those with subclinical dysfunction or overt heart failure (HF).
Materials and Methods
Patients
We studied 60 subjects: 20 healthy controls aged 18 to 30 years, 20 patients with type 2 diabetes, and 20 patients with severe symptomatic HF treated by resynchronization therapy. All subjects participated in clinical research studies in the Wales Heart Research Institute, which were approved by the local research ethics committee; details of the research protocols used in these studies have been published. The normal patients were the first 20 control subjects recruited for an ongoing study that evaluates subclinical myocardial and vascular dysfunction in young patients with type 1 diabetes. All patients gave written informed consent. The characteristics of the study groups are shown in Table 1 .
| P value (ANOVA with Bonferroni correction) | ||||||
|---|---|---|---|---|---|---|
| Normal (n = 20) | Type 2 DM (n = 20) | HF (n = 20) | Normal vs type 2 DM | Normal vs HF | Type 2 DM vs HF | |
| Demographic data | ||||||
| Age (y) | 25.7 ± 2.9 | 64.1 ± 8.5 | 64.6 ± 7.7 | <.001 | <.001 | 1.00 |
| Male | 50% | 70% | 85% | .51 | .054 | .90 |
| IHD | – | 10.5% | 35% | – | – | .151 a |
| Hypertension | – | 78.9% | 15% | – | – | <.001 a |
| Diabetes | – | 100% | 20% | – | – | <.001 a |
| Treatment | ||||||
| Aspirin | – | 63% | 20% | – | – | .003 a |
| Beta-blockers | – | 31.2% | 75% | – | – | .017 a |
| ACE-I | – | 75% | 100% | – | – | .101 a |
| Statin | – | 73.4% | 65% | – | – | .81 a |
| Echocardiographic data | ||||||
| Frames per second | 174.6 ± 19.1 | 160.7 ± 26.0 | 184.9 ± 24.7 | .20 | .51 | .006 |
| Heart rate (beats/min) | 65.1 ± 12.5 | 66.2 ± 10.5 | 70.2 ± 14.1 | 1.00 | .61 | .94 |
| LA (mm) | 35.7 ± 4.5 | 39.3 ± 6.9 | 45.1 ± 7.5 | .25 | <.001 | .02 |
| IVS (mm) | 8.4 ± 1.4 | 12.4 ± 3.2 | 10.8 ± 2.2 | <.001 | .009 | .10 |
| PW (mm) | 8.6 ± 1.4 | 11 ± 2.8 | 10.9 ± 2.2 | .003 | .006 | 1.0 |
| LVEDD (mm) | 48.9 ± 4.4 | 45.2 ± 6.3 | 62.5 ± 8.5 | .26 | <.001 | <.001 |
| LVESD (mm) | 33.8 ± 4.2 | 32.1 ± 4.8 | 53.7 ± 9.5 | 1.0 | <.001 | <.001 |
| LVEDV (mL) | 102.8 ± 21.6 | 81.7 ± 27.6 | 195 ± 78.9 | .58 | <.001 | <.001 |
| LVESV (mL) | 41.6 ± 9.5 | 35.2 ± 14.5 | 135.8 ± 62.6 | 1.0 | <.001 | <.001 |
| LVEF (%) | 59.5 ± 4.7 | 56.2 ± 7.0 | 31.5 ± 9.2 | .46 | <.001 | <.001 |
| Peak mitral annular velocities during ejection (cm/s) | ||||||
| Septal | 6.6 ± 1.1 | 5.5 ± 1.3 | 3.3 ± 1.5 | .033 | <.001 | <.001 |
| Inferior | 6.9 ± 1.1 | 5.8 ± 1.5 | 3.7 ± 1.6 | <.001 | <.001 | <.001 |
| Posterior | 7.4 ± 1.6 | 5.9 ± 1.8 | 4.1 ± 1.5 | .013 | <.001 | .002 |
| Lateral | 8.0 ± 1.3 | 5.8 ± 1.3 | 3.7 ± 1.5 | <.001 | <.001 | <.001 |
| Anterior | 7.3 ± 1.5 | 5.0 ± 1.8 | 3.4 ± 1.4 | <.001 | <.001 | <.001 |
| Anterior septal | 6.6 ± 1.3 | 4.4 ± 1.5 | 3.5 ± 1.8 | <.001 | <.001 | .28 |
| RV free wall | 9.9 ± 2.6 | 10.3 ± 2.8 | 6.8 ± 2.3 | 1.0 | <.001 | <.001 |
a P values were computed between type 2 DM and HF groups, excluding normal controls.
Echocardiographic Studies
Echocardiographic studies were acquired using a System V or Vivid 7 machine (General Electric, Horten, Norway) equipped with a 2.5-MHz probe. Color tissue velocity loops were stored for off-line analysis; images were optimized to obtain frame rates exceeding 100 s −1 , as recommended.
All measurements were performed by off-line analysis using commercially available software (Echopac version 5.2.0, General Electric). LV dimensions and function were measured according to current recommendations: end-diastolic thicknesses of the ventricular septum and the LV posterior wall, end-diastolic and end-systolic LV diameters, end-systolic left atrial diameter, and LV ejection fraction by the biplane Simpson rule.
Longitudinal IVA of the LV was analyzed from apical 4-chamber, 2-chamber, and long-axis images. In each “wall,” IVA was measured by positioning the cursor during end systole over the annulus and then in the basal third of the basal and mid-ventricular segments. Each measurement was made by dividing the first peak positive velocity during isovolumic contraction by the interval at which it occurred after the onset of the isovolumic velocity signal (at the zero crossing) ( Figure 1 ); thus all the measurements represent mean IVA rather than instantaneous peak IVA. Myocardial velocities were also measured off-line from the stored apical loops, during both isovolumic contraction and systolic ejection, in both the LV and the RV. All measurements were made if possible as the arithmetic mean of values obtained from 3 consecutive beats; occasional beats were discarded if the signals were too noisy to be analyzed.
Feasibility and Reproducibility Studies
The feasibility of measuring IVA was assessed from the first set of measurements made in all 60 subjects by the first observer, giving a total of 1,500 segments. Intraobserver reproducibility was assessed in all 60 subjects between 2 measurements made by the same observer on different days. Interobserver reproducibility was assessed between pairs of 3 independent observers who each measured 30 studies (comprising a random sample of 10 from each group). Before these independent analyses were conducted, joint review of examples was used to develop a consensus for the measurement of signals that did not conform to standard expected patterns; this consensus was then applied independently by the 3 observers, as follows.
Sampling sites
A sample size of 6/6 mm was placed over the annulus, midway between its maximal displacements toward the apex and the atrium (ie, in mid-excursion). Each myocardial wall was divided into 3 segments of equal length between the mitral or tricuspid annulus and the apex (ie, using standard echocardiographic segmentation). The sample volume (cursor) was placed at the basal part of the basal and mid-ventricular segments, during maximal systolic apical displacement ( Supplementary Figure 1 ). Minor adjustment of the sample position was allowed to give the clearest IVA signal with the least variation between consecutive beats.
IVA measurement
IVA was measured as the slope of the first positive deflection after the onset of systole (as defined by the surface electrocardiogram), during isovolumic contraction, from the intersection of the velocity trace with zero velocity, to the corresponding peak isovolumic velocity. The IVA was only measured if there was a clear positive deflection followed by a negative deflection, during isovolumic contraction, with an identifiable peak ( Figure 1 ). The following rules were also applied:
- •
if there was a single peak during isovolumic contraction, that peak was used to measure IVA;
- •
if there was no clear peak, the IVA was not measured;
- •
if the isovolumic tracing consisted of a single convex line, the maximal peak of that line was used to measure IVA;
- •
if the isovolumic tracing had a peak plateau, the first point of that plateau was used to measure IVA; and
- •
if there were 2 or more distinctive peaks of equal or different velocities during isovolumic contraction, the first peak was used to measure IVA.
Statistical Analysis
Quantitative data are presented as mean ± standard deviation (SD). Correlations between independent variables are reported using the Pearson correlation coefficient. Comparisons between categoric values were computed by chi-square test. Comparisons between multiple parametric values were computed by analysis of variance, applying the Bonferroni correction. A P value < .05 was considered statistically significant.
Reproducibility is expressed as the CV, unless stated otherwise. The CV was calculated using the formula: CV = SD/(arithmetic mean of measurements) × 100, where SD is the standard deviation of the measurement error associated with a single measurement, calculated as the SD of residuals (measurement 1 – measurement 2) divided by √2.
Interobserver variability was calculated between each pair of observers, as the CV, and then a pooled CV was obtained by averaging the 3 values; results are reported as mean CV ± SD. Each pair of observations was also compared using a Bland–Altman analysis to estimate systematic differences between observers; pooled results are reported as the mean difference in measurements observed in the 3 comparisons between different pairs of observers.
Results
Long-axis myocardial acceleration toward the apex during isovolumic contraction is highest when measured at the mitral annulus. In the 6 echocardiographic “walls” of the left ventricle and the RV free wall, acceleration decreased from the level of the annulus to the mid-ventricular segments by 24% to 43%. Results for the 3 separate study groups are shown in Table 2 , and the mean values for all 60 subjects are illustrated in Figure 2 . There was little regional variation in longitudinal IVA between LV myocardial segments at the same level (annular, basal, or mid), apart from IVA at the annular anterior septum, which was on average 29% lower than IVA in the other LV annular sites, and IVA at the mid-anterior septum, which was 28% lower than IVA at the mid-septal site ( Figure 2 ). IVA measured in the RV free wall was on average 52% higher than in LV segments at the same level.
| IVA (m/s 2 ) | P value (ANOVA with Bonferroni correction) | ||||||
|---|---|---|---|---|---|---|---|
| Normal | Type 2 DM | HF | Normal vs type 2 DM | Normal vs HF | Type 2 DM vs HF | ||
| Septum | a | 1.6 ± 0.7 | 1.5 ± 0.8 | 1.0 ± 0.6 | 1.0 | .035 | .064 |
| b | 1.1 ± 0.5 | 1.0 ± 0.6 | 0.9 ± 0.4 | 1.0 | .24 | .96 | |
| m | 0.8 ± 0.3 | 0.8 ± 0.4 | 0.9 ±0.6 | 1.0 | 1.0 | 1.0 | |
| Inferior | a | 1.3 ± 0.5 | 1.4 ± 0.8 | 1.1 ± 0.5 | 1.0 | .92 | .23 |
| b | 1.2 ± 0.6 | 1.2 ± 0.7 | 0.9 ± 0.6 | 1.0 | .79 | .69 | |
| m | 0.7 ± 0.3 | 0.9 ± 0.5 | 0.7 ± 0.4 | .14 | 1.0 | .55 | |
| Posterior | a | 0.9 ± 0.4 | 1.4 ± 0.7 | 1.1 ± 0.6 | .046 | .85 | .53 |
| b | 0.9 ± 0.5 | 1.1 ± 0.7 | 1.0 ± 0.7 | .79 | 1.0 | 1.0 | |
| m | 0.9 ± 0.6 | 0.9 ± 0.6 | 0.8 ± 0.5 | 1.0 | 1.0 | .94 | |
| Lateral | a | 1.3 ± 0.6 | 1.3 ± 0.6 | 1.0 ± 0.6 | 1.0 | .29 | .23 |
| b | 1.2 ± 0.7 | 1.0 ± 0.6 | 0.8 ± 0.7 | 1.0 | .24 | 1.0 | |
| m | 0.9 ± 0.7 | 0.8 ± 0.5 | 0.8 ± 0.7 | 1.0 | 1.0 | 1.0 | |
| Anterior | a | 1.6 ± 0.6 | 1.4 ± 0.8 | 0.9 ± 0.6 | .98 | .006 | .08 |
| b | 1.3 ± 0.6 | 1.2 ± 0.6 | 1.0 ± 0.6 | 1.0 | .39 | 1.0 | |
| m | 0.8 ± 0.3 | 0.8 ± 0.5 | 0.7 ± 0.5 | 1.0 | 1.0 | 1.0 | |
| Anterior septum | a | 1.0 ± 0.4 | 0.8 ± 0.5 | 0.9 ± 0.7 | .60 | 1.0 | 1.0 |
| b | 0.9 ± 0.3 | 0.7 ± 0.5 | 0.8 ± 0.4 | .56 | 1.0 | 1.0 | |
| m | 0.6 ± 0.3 | 0.6 ± 0.4 | 0.6 ± 0.4 | 1.0 | 1.0 | 1.0 | |
| Right ventricle | a | 1.8 ± 0.5 | 1.8 ± 0.8 | 1.4 ± 0.8 | 1.0 | .45 | .24 |
| b | 1.6 ± 0.7 | 1.9 ± 0.5 | 1.3 ± 0.9 | .95 | .47 | .08 | |
| m | 1.2 ± 0.6 | 1.4 ± 0.8 | 1.3 ± 1.0 | 1.0 | 1.0 | 1.0 | |
In the healthy young control subjects, the mean value and normal range (95% confidence intervals) for IVA of the medial mitral annulus was 1.6 m/s 2 (1.2-1.9 m/s 2 ). At the lateral tricuspid annulus, the normal value was 1.8 m/s 2 (1.5-2.0 m/s 2 ).
The only 2 sites where IVA identified overt myocardial dysfunction in patients with HF compared with healthy controls were the medial mitral annulus (annular septum) and the anterior mitral annulus ( Table 2 ). The comparison between subjects with HF and subjects with diabetes at these sites did not reach significance ( P = .064, and P = .08, respectively), and we were unable to identify any other sites where IVA was able to distinguish overt myocardial dysfunction in the population with HF from the subclinical disease present in the diabetic cohort. Counterintuitively, IVA at the posterior mitral annulus was higher in diabetic patients than in normal controls ( P = .046).
Unlike IVA, peak mitral annular velocities during ejection were able to identify each group from one another ( Table 1 ).
IVA correlated with the peak positive velocity during isovolumic contraction (in all segments, R > 0.78 and P < .0001). In the diabetic and HF groups, IVA also correlated with peak myocardial systolic velocities during ejection (R = 0.46-0.77; P < .01). IVA was weakly correlated with global LV ejection fraction but only when the relationship was tested for all subjects together (N = 60, R = 0.28, P = .031). There was no correlation between IVA and LV volumes.
It was possible to estimate ±dp/dt from continuous-wave Doppler profiles of mitral regurgitation in only 13 subjects (11 with HF and 2 with diabetes). In these 13 subjects taken together, IVA did not correlate with ±dp/dt, but the 2 diabetic subjects had widely discrepant estimates of ±dp/dt (935 and 3732 mm Hg/s). In the 11 subjects from the HF group, ±dp/dt was 754 ± 159 mm Hg/s, and there were trends for this to correlate with IVA measured at the medial mitral annulus (R = 0.65, P = .09), the basal septum (R = 0.61, P = .08), the inferior mitral annulus (R = 0.55, P = .09), and the basal lateral LV segment (R = 0.55, P = .08). IVA correlated best with ±dp/dt when it was measured at the anterior annulus and in the basal and mid-anterior LV segments (R = 0.63-0.78, all P < .04).
Feasibility
Measurement of IVA was feasible in 1340 of the 1500 segments analyzed (89%) for the whole group. Feasibility was 97% in the normal subjects, 89% in the type 2 diabetic patients, and 82% in the patients with HF ( P < .001). In the normal group, the feasibility at all sites was greater than 85%. In the HF group, no segment showed a feasibility of 100%. The principal factor contributing to reduced feasibility was inability to define a clear positive velocity peak during isovolumic systole.
Reproducibility
Intraobserver reproducibility was calculated for all the subjects taken together (N = 60). CV for repeated measurements of longitudinal IVA ranged from 11% (at the basal inferior segment) to 32% (in the mid-RV segment) ( Table 3 ); the average from all 1340 segments was 20%. CV at the inferior annulus, mid-ventricular inferior segment, and medial mitral annulus were also less than 15%. Reproducibility consistently decreased from annular to mid-ventricular sites. All measurements from the anterior septum had low reproducibility (CVs 24%-28%).
