Development of a multiparametric score to predict left ventricular remodelling and prognosis after cardiac...

Optimal delivery of CRT requires appropriate patient selection and device implantation. Echocardiographic predictors of CRT response individually appear to enhance patient selection, but do not fully reflect the complex underlying myocardial dysfunction. We hypothesized that a multiparametric approach would offer greater predictive value and sought to derive a score incorporating baseline characteristics including: dyssynchrony, LV function, and LV lead position.

Data were analysed from 294 patients undergoing CRT between June 2008 and December 2012. All patients were in sinus rhythm with QRS >120 ms, NYHA class II-IV, and LVEF <35%. Detailed clinical assessment including echocardiography was completed at baseline and 6 months after CRT. Response was defined as a ≥15% reduction in LV end-systolic volume. Dyssynchrony (interventricular delay and radial strain delay), global longitudinal strain, and LV lead position were independent predictors of LV remodelling and were used to derive a predictive score which correlated with reduction in LV volume (r = - 0.5, P < 0.001) and was higher with QRS >150 ms and non-ischaemic aetiology. A cut-off score <0.6 offered the highest specificity and positive predictive value (100%) to determine non-response. A score >3.28 offered high specificity (specificity 86%, sensitivity 70%) to predict response. Survival proportion at longer term follow-up was low (21%) in the group with predictive score <0.6.

A multiparametric strategy, which defines anticipated probability of response to CRT, offers potential to predict non-responders with poor long-term survival following CRT. The value of this approach in avoiding unnecessary device implantation with potential for harm requires validation in large multicentre studies.