Predicting scores for left ventricular dysfunction in Duchenne muscular dystrophy

Background: The aim of this study was to predict left ventricular (LV) dysfunction and the timing to perform echocardiography in patients with Duchenne muscular dystrophy (DMD). We developed a scoring system using clinical parameters and examined its efficacy. It is indispensable to utilize echocardiogram for evaluating myocardial damage of DMD patients, but there is no established guideline for determining the clinical conditions which require echocardiographic examination. Methods: We retrospectively analyzed 86 patients with DMD who were treated in Kobe University Hospital from 2007 to 2009. The multiple logistic regression analysis on routine clinical data was performed to identify parameters that can find abnormal LV contraction, and to develop a weighted scoring system. Echocardiogram was performed as the gold standard for detecting LV dysfunction. Results: Four parameters were associated with abnormal LV contraction: (i) brain natriuretic peptide (BNP); (ii) creatine kinase; (iii) scoliosis; and (iv) body surface area. When BNP was used as the only predictor to evaluate LV systolic dysfunction, sensitivity and specificity were 36.4% and 92.1%, respectively. In contrast, abnormal LV contraction was detected in high accuracy (sensitivity: 95.5%; specificity: 68.3%) when we used a two-step scoring system in which BNP was combined with the other three factors, raising the sensitivity compared to using BNP levels as the single parameter (P= 0.008). Conclusion: Our scoring system detects the early heart dysfunction of DMD patients, especially when BNP level is not elevated. This system is useful to determine the timing for echocardiographic examination and consulting cardiologists.