Serum Ferritin and Liver Iron Burden by MRI Predict Cardiac Iron Overload in Thalassemia Major.

Abstract

Iron overload remains a substantial challenge in thalassemia major. Classically, liver biopsy and serum ferritin have been used as surrogates for cardiac risk. However, MRI is gaining increasing acceptance for noninvasive assessment of liver and cardiac iron burden. The purpose of this study was to compare the prognostic value of liver iron and long-term trends in serum ferritin for predicting cardiac iron burden. Sustained serum ferritin concentration above 2500 ng/ml due to ineffective chelation is an established prognostic factor for cardiac disease-free survival. We studied 18 patients with thalassemia major from a single U.S. institution. Iron assessments with MRI of the liver and heart were compared to cardiac function, average serum ferritin over the past 5 years, age, years of transfusion and, in a limited subset, liver biopsy (bx) obtained within 6 months of MRI studies. The mean age of the population was 25.2 yrs (range 12–44). Iron chelation was achieved with deferasirox (n=14), deferoxamine (n=3) or deferriprone (n=1). Cardiac and liver MRI R2* using a multiple gradient echo sequence were performed on a single 1.5T GE scanner. Hepatic iron concentration (HIC) values were predicted from liver R2* values using the method of Wood et al. HIC values by MRI ranged from 2.4–29 mg Fe/g dry wt. The overall agreement of HIC using this MRI R2* method with mean ferritin values was marginal (r=0.477), however a much stronger correlation was demonstrated when subjects with outlying ferritin values >4500 (n=3) were excluded (r=0.747, p=0.001). Using a t-test for independent samples, we found that the mean HIC among patients who maintained ferritin values >2500 was 18.0, while patients with average ferritins <2500 had a mean HIC of 7.4 (p=0.003). There was also a striking inverse correlation between average ferritin and cardiac T2*. Patients with 5 yr ferritin averages >2500 had a mean T2* value of 9.9 msec, while patients with mean values <2500 had a median T2* of 37 msec (p=<0.001). We found no correlation of age to cardiac T2*and were unable to evaluate the relationship of T2* to ventricular function in this relatively small cohort. Previous studies have found marked discordance between liver and cardiac iron measurements. The correlation of liver MRI R2* and cardiac MRI T2* in our population was modest (r= −0.60, p=0.009). A subset of patients (n=8) had liver bx performed within 9 ± 8 weeks of the MRI. As noted previously by others, we found that the HIC from liver bx had a very strong linear correlation with mean serum ferritin (r=0.869, p=0.005). As an additional validation of MRI R2*, standard liver bx results were comparable to HIC by MRI, but only for bx values <25 (r=0.992, p=0.001).Conclusion: New technologies underscore that poor compliance with chelation treatment based on serum ferritin remains an important indicator of iron burden and risk of complications related to iron induced organ failure. Estimations of hepatic iron burden by MRI were comparable to results of liver bx for mild to moderate overload, however, severe overload correlated less well. This likely reflects the technical limitations of our current R2* implementation and must be considered in utilizing MRI in heavily burdened subjects. High ferritin values likely due to poor chelation compliance correlates strongly with increased myocardial iron based on lower T2* values. A larger longitudinal study will be needed to determine if low T2* predicts cardiac disease-free survival.