Cardiac complications of transfusional iron overload are well documented in various inherited anemias. In regularly transfused MDS, the deleterious role of iron overload on cardiac disease is more disputed, due in particular to frequent concomitant causes of cardiac failure. Cardiac MRI T2* allows accurate and specific measurement of iron content.
We prospectively evaluated in 4 centers of the GFM by standardized and transferable MRI methods both cardiac T2* according to Anderson (Eur Heart J. 2001Dec;22(23):2171-9) and liver iron content (LIC) according to Gandon (Lancet. 2004 Jan 31;363(9406):357-62), as well as cardiac function by routine echocardiography or MRI in regularly transfused MDS patients.
From Dec 2005 to March 2010, 73 patients (pts) were included (14 of them had more than one MRI evaluation over time): 38 M/35F, Median age 68 (24-86); WHO : RA=5, RARS=33, RMCD-RS=3, RMCD=1, RAEB1=9, RAEB2=5, RAEB-T/AML=1, 5q- syndrome=8 and unclassified=8; Karyotype: fav n=50, Int n=9, unfav n=4, failure n=10; IPSS: low n=29, Int-1 n=28, Int-2 n=5 and High n=1, unknown n=10. Median interval from MDS diagnosis and MRI T2* assessment was 49 months (range 0–324). Median serum ferritin at MRI assessment was 1750 ng/ml (range 282–7339) and 54/73 pts were on chelation therapy (CT) (median duration of CT prior to first MRI: 18 months, range 1–125). 37/73 pts had cardiac symptoms and 28 were on cardiac therapy. At first MRI T2* analysis, the median number of RBC units transfused was 68 (range 5–574). Median LIC was 330 micromoles/g/dw (range 40–908). Median Cardiac T2* was 27 ms (range 6–74). 14/73 pts had cardiac iron overload defined by MRI T2* ≤20 ms (19%) and among them 3/73 (4%) had severe cardiac iron overload (T2*≤ 10 ms). LVEF was below normal (55%) in 13/59 cases evaluated. A correlation was found between cardiac T2*and the number of RBC units transfused (Pearson correlation =-0.342, p=0.004) but not with LIC (p= 0.65) and serum ferritin (p=0.21). Cardiac overload was seen in 1/19 (5.5%) pts transfused <50 RBC units, 4/37(12.1%) pts transfused 50–150 units, 9/17 (52.9%) pts transfused >150 units (p= 0.0005). Those 3 pt subgroups also differed in median LIC (μmoles/g/dw) (<50 units= 250, 50–150 units=340, > 150 units=414) (p=0.044 Kruskall-Wallis' test), but not significantly in serum ferritin (p= 0.085).
No significant correlation was found between decreased LVEF (< 55%) and cardiac T2* <20 ms (p=0.5), or T2*≤10 ms (p=0.23). In particular, 5/13 pts (38%) with LVEF <55% had T2*<20ms, vs versus 8/46 pts (17%) with LVEF >55% (p= 0.13). However, 1/14, 0/30 and 3/12 pts having received <50, 50–150 and > 150 RBC units had severe cardiac failure (ie LVEF≤35%)(p=0.012). 3/4 pts with severe cardiac failure had T2*< 20ms,compared to 8/54 pts without severe cardiac failure (p=0.023).
14 pts had another cardiac MRI 6 to 34 months (median 18) after the first. All were on CT and had received a median of 60 and 214 PRBC units at first and last MRI, resp. Median Cardiac T2* was 21.6 ms (range 8.5–35.3) and. 28 ms (range 6.4–41) at last and at first assessment, respectively (p=0.3)
Moderate and severe post transfusional cardiac iron overload was seen in 19% and 4% of regularly transfused MDS, respectively. The level of cardiac iron overload was well correlated to the number of RBC transfused. The impact of cardiac overload on LVEF was unclear except in pts with severe cardiac impairment (LVEF <35%), possibly suggesting that iron overload is only one of the factors responsible for cardiac disease in many of those elderly patients.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.