Abstract
Abstract 2013
Poster Board I-1035
Background:
Despite recent advances in their treatment, cardiac morbidity remains a significant concern in patients with transfusion dependent thalassemia. Cardiac magnetic resonance (CMR) has been introduced as the tool of preference in evaluating cardiac iron load in chronically-transfused patients. Data on the relationship of the degree of cardiac iron load, as assessed by CMR, and the occurrence of cardiac events are limited. Furthermore, the value of CMR in predicting cardiac event seems to change with the use of more cardioprotective iron chelating agents.
Patients and Methods:
Cardiac iron load was estimated by measuring the myocardial relaxation parameter T2* by CMR. Data from the first-ever CMR of 318 patients with transfusion-dependent thalassemia followed in a single institution were included in this cross-sectional study. A patient was characterized as having cardiac disease (CD) if he/she fulfilled at least one of the following criteria: Left Ventricular Shortening Fraction (LVSF) <30% as assessed by echocardiogram, arrhythmia or requiring therapy with cardiac medications for clinically evident heart dysfunction. The predictive value of CMR was estimated by reviewing the cardiac events in CD-free patients, which occurred in the period of time between the initial CMR and December 31, 2008.
The mean T2* between CD-free and CD patients was compared by Student t- test, the ratio was compared by Fisher's exact test and the relative risks (RR) were estimated by logistic regression. Receiver operating characteristic analysis (ROC) was used to evaluate CMR's ability to discriminate between CD and CD-free patients.
Results:
At the time of their first CMR, 77 patients were characterized as having CD. Their mean T2* was significantly lower than that of the 241 CD-free patients (11.0 ± 9.4 vs 23.1 ± 11.7 ms, p< 0.001). The percentages of CD-free patients increased along with the T2* (53% with T2*<8ms, 68% with T2* between 8-14ms, 79% with T2* between 14-20 ms and 92% with T2*>20, p<0.001). The RR of having CD for patients with T2*<8ms vs 8-14ms, 14-20ms and >20ms were 1.9(p=0.07), 3.4 (p=0.006) and 10 (p<0.001 ), respectively. Similarly, RR for T2* between 8-14ms vs >20 was 5 (p<0.001) and between 14-20 ms vs >20 was 3 (p<0.018). With respect to the capability of T2* to discriminate between CD and CD-free patients, the ROC analysis estimated an area under the curve equal to 0.79 (95% C. I. 0.73-0.85). At T2* cut-off values of 20, 17, 14 and 10 ms, the %'s for sensitivity/specificity are 85.3/56, 77.9/62.8, 75/70 and 63.2/79.6 respectively. The value of CMR in predicting cardiac events in CD-free patient according to the initial T2* was 2.9/100 patient-years for T2*<8ms, 0.9/100 patient/years for T2* between 8-14ms or 14-20ms, decreasing to 0.4/100 patient-year for T2*>20ms.
Conclusions:
CMR is a sensitive tool in determining the risk for potential cardiac dysfunction in chronically-transfused patients. Its predictive value for occurrence of cardiac events, even in patients with severe cardiac iron load, seems to be limited in patients that change iron chelation therapy. The predictive value of CMR levels could better be assessed if patients are monitored prospectively. The implementation of CMR in the regular follow-up of the patients coincided with the changes in the chelation therapy, namely the introduction of deferiprone and deferasirox, which changed dramatically the cardiac morbidity and mortality of these patients. These changes explained at least in part the low predictive value of the CMR.
Disclosures:
Kattamis: Novartis: Consultancy, Honoraria, Speakers Bureau. Berdoukas:ApoPharma Inc: Consultancy, Honoraria, Speakers Bureau; Novartis: Confidentiality agreement for development of ICL 670 and attendance at Scientific Sessions sponsored by Novartis. Ladis:Novartis: Honoraria; ApoPharma Inc: Honoraria.
Cardiac magnetic resonance in transfusion dependent thalassaemia: assessment of iron load and relationship to left ventricular ejection fraction
