Optimizing the Management of Iron Overload in Patients with Hemoglobin Disorders in Centers with Limited Access to Magnetic Resonance Imaging

Background:Transfusion therapy for severe chronic anemias is a major cause of iron overload. Patients with hemoglobin disorders are often transfused, either chronically or occasionally, and if iron overload is suspected, it is currently recommended they are subjected to so-called "iron studies" and determination of liver iron content (LIC) and heart T2* by magnetic resonance imaging (MRI). It remains common practice to request both laboratory and imaging tests, even for patients already undergoing iron chelation, despite the unreliability of laboratory markers such as ferritin in the presence of concurrent inflammatory conditions, which do not affect results from MRI scans. Since MRI is not widely available in most underdeveloped and developing countries, we aimed to assess the contribution of laboratory tests to the management of patients with hemoglobin disorders with iron overload by analyzing a cohort of patients who managed to undergo MRI during a period of limited access to MRI scans in our center.

Methods: In this single-center, retrospective analysis, data were collected from digital MRI, laboratory information system, and electronic chart databases, upon approval of the local ethics board. Since access to MRI was limited during the period of time studied, physicians used the following criteria to indicate MRI scans to define clinical suspicion of iron overload: patient already undergoing chelation therapy based on previous laboratory tests and clinical assessment, current or previous chronic transfusion therapy, or elevated ferritin with transferrin saturation (TSAT) above 45%.

Results and Discussion: From April 2014 to September 2016, 27 patients (17 with sickle cell disease - SCD, 10 with beta thalassemia - BT), median age 35 years (15-62), 9 male (33%), were evaluated at 34 different occasions with MRI and conventional iron studies. Patients with BT had significantly higher levels of serum iron and TSAT than patients with SCD (BT median 266ug/dL, range 93-328 and 0.94, range 0.41-1.00; SCD median 157ug/dL, range 58-264, and 0.64, range 0.21-1.00, P=0.004 and P=0.0138), which reflects a higher percentage of BT patients on chronic transfusions (8/11 vs. 6/17) and the higher transfusion volume for BT in comparison to what is needed for exchange transfusions in SCD. Four patients had normal MRI scans: 1 with SCD not in transfusion, 1 with SCD that had been treated with iron chelation, and 2 BT major in chronic transfusion and optimal response to iron chelation (ferritin <500ng/mL). Only one SCD and one BT patient were diagnosed with cardiac iron overload, with a T2* of 7.5ms and 2.95ms, respectively, in patients with extreme hyperferritinemia (11,567 and 7,090ng/mL) and low compliance to iron chelation. MRI liver iron concentration correlated positively with serum ferritin and total iron binding capacity (TIBC) measurements (r=0.67, 95% CI 0.41-0.83, P <0.0001, and r=-0.49, 95% CI -0.72 to -0.17, P <0.004, respectively) but no serum biomarkers correlated with myocardial iron concentrations. Receiver operating characteristic (ROC) curve analysis showed that ferritin levels performed best in our population as predictor of any increase of LIC above 2mg/g (area under curve 0.9753, P=0.0003), with a cutoff value of 731.6ng/mL with a sensitivity of 96.3% and specificity of 100% to discriminate for the presence of any increase in hepatic iron. In addition, a ferritin cutoff value of 1649ng/mL yielded 76.2% sensitivity and 91.7% specificity to detect a LIC above 5mg/g, comprising 17 (50%) of the MRI scans. Only one patient in this group had less than moderate liver iron overload, but still with an abnormal LIC of 3.5mg/g.

Conclusions: Our data suggest that conventional iron studies cannot predict cardiac siderosis, but earlier evaluation with MRI may be warranted in the presence of clinical risk factors such as history of chronic transfusion along with low compliance to iron chelation, independently from laboratory tests. In settings with limited access to MRI, ferritin levels may be useful in pre-screening for hepatic iron overload, so postponing probable negative scans for patients undergoing iron chelation with optimal response or probable positive scans for those with certain levels of hyperferritinemia (e.g.>1649mg/dL) could be appropriate criteria to manage referrals to MRI scans and help prioritize patients with undetermined indication of iron chelation.