Pancreatic Iron and Fat Infiltration Are Diabetogenic Risk Factors In Patients With Iron Overload

Chronically transfused patients and patients with hereditary hemochromatosis (HHC) are affected by complications due to excessive body iron accumulation. Therefore, early detection of tissue iron accumulation is mandatory for effective treatment. Besides serum ferritin (SF) as the traditional blood parameter indicating iron overload, the assessment of body iron stores by non-invasive magnetic resonance imaging (MRI) using R2- (T2) or R2*- (T2*) relaxometry is established in clinical routine. Especially the comparison of tissue iron burden and iron loading patterns is a current field of research in iron overload.

Although the pancreas seems to play a major role in the endocrinological pathology of patients with siderosis, there are only few studies analyzing pancreatic iron content referring to the head, body and tail region of the pancreas.

A total of 69 patients (age 9 - 78 years, 31 females) with ß-Thalassemia major (TM: n = 42; 16 splenectomized), Diamond-Blackfan anemia (DBA: n = 7; 1 splenectomized), hereditary hemochromatosis (HHC: n = 11) and iron overload from chronic blood transfusions (n = 9; sideroblastic anemia, aplastic anemia, myelodysplastic syndrome, leukemia post bone marrow transplantation) were scanned between 2007 and 2012 as part of their regular heart and liver iron monitoring by MRI, respectively. 10 healthy subjects (3 females, age 23 - 68 years), voluntarily served as controls.

The study was approved by the local ethics committee and all subjects gave their written informed consent.

For measuring pancreatic R2*, a stack of 4-8 of slices (thickness = 5.5 mm, no gap, inplane resolution 1.25x1.25 mm²) was selected, covering the whole pancreas.

Pancreatic signal intensities were averaged from three different ROIs positioned in the tail, body and head of the pancreas with sufficient distance to the gland boundaries and excluding vascular structures. Only clearly identified and demarcated pancreas tissue was considered.

Since fat infiltration is a common problem in the pancreas, especially with phase dependent GRE times, we applied a water/fat separation technique to the in-phase and out-of-phase signal intensities as described by Wehrli et al.

For patients, the median hepatic R2* rate, pancreatic R2* rate and ferritin level were 261 s^-1 (range 42 - 2326 s^-1), 130 s^-1 (range 23 – 1192 s^-1) and 1445 µg/ml (range 62 – 21557 µg/ml), respectively.

In 70% of all patients, fatty infiltration of the pancreas was above the range of controls (> 10%), with highest apparent fat content (aFC) in the pancreas of TM patients, especially, in TM patients with diabetes mellitus (median aFC = 29.1%). The highest aFC was found in the pancreatic tail (P = 0.2). Fat infiltration correlated with pancreatic R2* (r_S = 0.67, P < 10^-4).

We compared R2* rates in the different pancreatic regions involving patient groups of different iron loading diseases. Patients showed significantly higher pancreatic R2* rates and apparent fat contents than controls. The pancreas tail revealed the highest R2* rates and fat contents so we hypothesize that iron uploading initially occurs in the pancreas tail, especially in TM patients. Fatty replacement and degeneration of the pancreas seems to be an important risk factor on top of pancreatic iron burden for the development of diabetes and should be further investigated in longitudinal studies.

No relevant conflicts of interest to declare.