Introduction:
Presence of subclinical left ventricle (LV) dysfunction may influence outcomes after percutaneous mitral balloon valvotomy (PMBV), but cannot be detected by conventional echocardiography.
Hypothesis:
Preprocedure LV global longitudinal strain (GLS) by speckle tracking TTE detects subclinical LV dysfunction and predicts long-term outcome after successful PMVB.
Methods:
Consecutive patients undergoing PMBV between 2003-2011 who had comprehensive baseline TTE were included. Patients with suboptimal immediate results (repeat PMBV or mitral surgery within first 6 months) and those with suboptimal TTE strain were excluded. LV strain was analyzed with commercial software (Syngo VVI), and GLS averaged from 18 segments. The primary end point was all-cause mortality. Secondary end point was the composite of death and mitral valve reintervention.
Results:
Of 110 patients screened, 29 (26%) had unsuitable strain imaging, and 8 (7%) had poor immediate results, leaving 73 in the final cohort. Median follow-up was 3.7 years (interquartile range 1.3-7.2). There were 10 deaths (14%) and 25 late mitral valve interventions (34%). Univariate predictors of death were GLS, age, EF, stroke volume, pulmonary artery pressure, mitral valve area and coronary artery disease. In a multivariate Cox model, GLS was the strongest predictor of death after adjustment for age [HR=1.40 (1.09-1.84), P=0.008]. For the composite endpoint, GLS, mitral valve gradient, and invasively mean left atrial pressure were univariate predictors of death or reintervention. At multivariate analysis, GLS was the sole predictor of events [HR=1.17 (1.01-1.34), P=0.03]. Event-free survival is shown in figure. Intra and inter observer intra-class correlation for strain analysis was 0.87 (0.66-0.96) and 0.84 (0.53-0.95), respectively.
Conclusions:
GLS is a powerful predictor of long-term outcome after PMVB and appears to have an incremental prognostic value over traditional parameters.