Impact of Length Indexing of Deformation in Echocardiographic Evaluation of Right Ventricular Function.

BACKGROUND

When assessing right ventricular (RV) function by echocardiography, some discordance between the deformational indices is predicted on the basis of the influence of RV length. RV free wall longitudinal strain (RVFWS) is relatively independent of RV length, whereas tricuspid annular plane systolic excursion (TAPSE) reflects the strain-length product. Systolic annular velocity (s'; distance over time) is also likely to be influenced by length. The aim of this study was to test the hypothesis that indexing TAPSE and s' to RV length would lead to better congruency with RVFWS.

METHODS

Two separate cohorts were identified from retrospective data: (1) subjects with normal cardiac function (n = 75) and (2) a cohort with high likelihood of potential RV dysfunction determined by the study indication of either pulmonary hypertension or pulmonary embolism (n = 50). RV functional indices of TAPSE, RV s', RVFWS, and fractional area change were verified and remeasured by an expert. Correlations and concordance maps between RVFWS and either TAPSE or RV s' were made with and without indexing the latter measurements to RV length. Predictive accuracy for detecting abnormal RVFWS were made using receiver operating characteristics analysis.

RESULTS

In normal subjects, indexing either TAPSE or RV s' to RV length led to an improvement in the correlation coefficient (from 0.59 to 0.68 for TAPSE, from 0.41 to 0.58 for RV s') and the variance (F statistic from 64.9 to 105.3 for TAPSE from 24.7 to 63.9 for RV s') for correlations with RVFWS. In all subjects, categorical concordance with RVFWS was improved by indexing TAPSE and s' to RV length primarily because of correction of underperformance to detect abnormal RVFWS in subjects with long RV length and better discrimination as normal for subjects with short RV length. Indexing to RV length improved the C statistic for detecting abnormal RVFWS for both TAPSE (0.80 vs 0.87, P = .03) and RV s' (0.65 vs 0.77, P = .002).

CONCLUSIONS

Indexing TAPSE and RV s' to RV length improves concordance of these deformational measurements with RVFWS and their ability to classify those with RV dysfunction according to RVFWS. Indexing TAPSE and RV s' to length is particularly effective for interpreting paradoxical information such as low TAPSE and s' in normal patients with short RV length and those with increased RV length who have normal TAPSE and s' values but other evidence of RV dysfunction.