To our knowledge, this is the first systematic review and meta-analysis on the use of TDI for the evaluation of CAD. Our results demonstrated significant differences in TDI measurements in those with and without CAD, including significant decreases in the maximum systolic velocity pre-stress, and maximum early diastolic velocity and maximum late diastolic velocity post-stress.
A correlation between decreased systolic velocity and reduced regional myocardial blood flow has also been demonstrated in animal models
 and a decreased maximum systolic velocity is highly sensitive and specific for detection of abnormal segmental myocardial motion
. In contrast, relatively little data is available specific to maximum early and late diastolic velocities. However, one study demonstrated that reduction of early diastolic velocity was more accurate than maximum systolic velocity for detection of CAD during dobutamine stress echocardiography
. A formal evaluation of the test characteristics (sensitivity, specificity, positive and negative predictive value) of TDI was not done because the studies reporting these were done in disparate patient population and used different indices of TDI
Our results suggest that decreases in maximal systolic velocity using TDI could be used to detect CAD in resting studies, especially in patients with poor two-dimensional endocardial definition. It could be incorporated into an echocardiogram based CAD diagnostic score, perhaps combined with wall motion score index, to improve the detection of CAD. Wall motion analysis is qualitative, and the addition of TDI adds a quantitative component in the detection of CAD, theoretically leading to improved intra- and inter-reader variability. Although not addressed by these data, abnormal velocities may also signal a more severe degree of CAD than suggested by wall motion analysis alone
Measurement of diastolic velocities could also be incorporated into stress testing algorithms, particularly dobutamine stress echocardiography, in which there is time to measure both two-dimensional and TDI parameters. In the ischemic cascade
, diastolic dysfunction occurs earlier than systolic dysfunction and thus measurement of diastolic velocities by TDI may prove more sensitive than wall motion analysis alone in the detection of CAD during stress testing.
However, the cutoff values for maximum systolic velocity and diastolic velocities which detect CAD are, as yet, unclear. Furthermore, TDI velocity can only be measured in one dimension and is significantly limited by angle dependence, complicating assessment of multiple wall segments (especially more apical segments). The absence of difference in the maximum systolic velocity post-stress with TDI may be a result of this limitation. Speckle tracking is a new technique that measures myocardial motion in every direction, thus overcoming the limitation of angle dependency and reliably quantifying myocardial ischemia in both infarcted and non-infarcted patients
[32, 33]. TDI velocities are also influenced by tethering and translation, movement of other adjacent structures and blood flow. Despite these limitations, the American Society of Echocardiography/European Association of Echocardiography guidelines acknowledge that TDI has major advantages like ready availability and quantitative evaluation through online measurement of velocities and time intervals
The conclusions of our meta-analysis are limited by the quality and relatively small number of included studies. Another limitation is the significant heterogeneity observed in the meta-analyses of maximum early and late diastolic velocity (pre-stress); and maximum systolic velocity and late diastolic velocity (post-stress). The heterogeneity could be the result of subtle differences in patient populations, segments in which TDI velocities were measured, models of the echocardiography machines used in the assessments, or type of tissue Doppler used (spectral Doppler, which measures the instantaneous velocity, versus color Doppler which measures the modal velocity). The methods of stress testing used (e.g., dipyridamole vs. dobutamine) to diagnose CAD were different in the included studies and present an intrinsic limitation in interpreting the analyses. Sensitivity analyses examining only those studies where CAD was evaluated using angiography showed a reduction in heterogeneity in some analyses (Table
3), suggesting that the method used to diagnosis CAD in the studies may have contributed to the observed heterogeneity. A third limitation is that data on head-to-head comparisons between TDI and other non-invasive imaging modalities or wall motion score analysis for diagnosing CAD are not available. It is thus unclear whether TDI has incremental value for diagnosing CAD when compared with resting (or stress) wall motion abnormalities or EKG abnormalities. However, resting wall motion abnormalities and EKG abnormalities are relatively insensitive for detection of CAD, so TDI may be helpful in improving the sensitivity of these tests. Another limitation is that coronary angiography was not performed in all patients across the studies. However, sensitivity analyses that excluded such studies showed similar meta-estimates. There may also be differences in the proportions of patients with variables like hypertension and left ventricular hypertrophy (LVH) across the studies included in this systematic review, and these differences could account for at least some of the heterogeneity observed across studies in the test characteristics of TDI for diagnosing CAD, as well as differences across studies in E/E' or E'/A' measurements.TDI would be low in patients with LVH or hypertension, potentially leading to false positives in patients with hypertension or LVH who do not have CAD. Thus, if the proportion of these patients differed across studies, then the test characteristics of TDI for diagnosing CAD might differ as a result. The proportion of patients with hypertension ranged from 21% to 58% across the included studies which could account for some of the observed heterogeneity in the test characteristics of TDI for diagnosing CAD. The proportion of patients with LVH was not reported in any study. A final limitation is that none of the above studies examined clinical outcomes, like symptomatic CAD.
In conclusion, our results suggest that TDI has a role in the evaluation of CAD. Future studies should evaluate the incremental value of TDI systolic and diastolic velocities over LV ejection fraction and two dimensional wall motion analysis in the detection of CAD and assessment of its severity.