Portal de Periódicos da CAPES

Optimizing the management of hereditary haemochromatosis: the value of MRI R2* quantification to predict and monitor body iron stores

2017; Wiley; Volume: 183; Issue: 3 Linguagem: Inglês

10.1111/bjh.14982

1365-2141

Manuela França, Luis Martí‐Bonmatí, Sara Silva, Carla Oliveíra, Ángel Alberich‐Bayarri, Filipa Vilas Boas, Helena Miranda, Graça Porto,

Trace Elements in Health

Hereditary haemochromatosis (HH) is a major cause of iron overload. Without therapeutic intervention, HH patients are at risk of severe tissue damage due to iron toxicity. HH-related morbidity and mortality can be effectively reduced by regular venesection (Niederau et al, 1985; Crosby, 1991), in the form of a course of weekly intensive phlebotomies (IPh) to deplete iron stores, followed by maintenance phlebotomies (MPh) to prevent iron re-accumulation (European Association for the Study of the Liver, 2010). In spite of being a well-accepted, efficient and safe procedure, phlebotomies may raise concerns regarding limitations of patients' daytime activities, decreasing their quality of life (Brissot et al, 2010). To manage patient expectation at diagnosis and optimize treatment compliance, it is important to predict the amount of mobilised iron and expected IPh duration, and to define reliable and safe endpoints for serum ferritin (SF) levels during the MPh. In this setting, low SF values (50–100 μg/l are usually advocated in clinical practice (European Association for the Study of the Liver, 2010), although normal SF (males: <300 μg/l; females: <200 μg/l) have also been proposed (Vanclooster et al, 2016). Whether iron re-accumulation differs according to different SF endpoints has not been reported to date. We have recently implemented a 3 Tesla multi-echo chemical shift-encoded gradient-echo magnetic resonance (MECSE-MR) sequence, with high accuracy to simultaneously quantify liver steatosis, assessed as proton density fat fraction (PDFF), and iron overload, assessed by R2* measurements (Martí-Bonmatí et al, 2011; França et al, 2017a,b). The same images also allow evaluation of the distribution of iron deposits in different abdominal tissues (França et al, 2017ab). This method might be useful for predicting mobilizable total body iron stores (TBIS) during IPh and for assessing liver iron re-accumulation in MPh. We recruited 48 phenotypically and genetically well-characterized HH patients (39 males, 15 females; age 19–80 years; 12 receiving IPh and 38 under MPh, 2 of which were also tested during IPh) and 21 apparently healthy controls (12 males, 9 females; age 20–77 years). TBIS were quantified in all 12 HH patients undergoing IPh, as previously described (Porto et al, 1997), considering the total amount of iron (in grams) removed by phlebotomies from the date of MR examination until iron depletion, defined as SF <50 μg/l. As patients were recruited at different IPh stages (3 of them had a second MR examination after total depletion), the calculated TBIS values ranged from 0 to 12·3 g. Notably, there was no significant correlation between initial SF values (retrieved from patients' data at diagnosis) and the TBIS quantified by phlebotomy (P = 0·578), possibly due to the generally poor specificity of high SF values, and the fact that, at diagnosis, values were distributed within a limited range of severe iron overload. Nevertheless, SF values determined at several time points during iron depletion (i.e. with a wide range of values from iron overload to iron depletion) were significantly correlated with the calculated TBIS at the same time points (RS = 0·916, P < 0·01). Therefore, SF appeared to be a good marker for monitoring iron depletion during IPh but could not predict TBIS at diagnosis. Using the data from the patients undergoing IPh, a stepwise regression analysis was performed considering TBIS as dependent variable, taking gender, age and all R2* tissue determinations (liver, pancreas, spleen and vertebra bone marrow) as independent variables (Table 1). Forward and backward selection methods were applied to select the most parsimonious model for predicting TBIS. The best regression model included two predictors: liver R2* (F-to-enter = 72·5553, P < 0·0001) and pancreas R2* (F-to-enter = 19·5504, P = 0·0008), with a coefficient of determination of R2 = 94·2% (P < 0·01), being expressed by the equation: TBIS = 0·006 × Liver R2* + 0·009 × Pancreas R2* − 1·064. This new TBIS variable was designated FELIPA, for iron (FErrum) in liver (LI) and pancreas (PA). Tissue iron quantification and distribution using R2* metrics was next assessed in the group of 38 patients undergoing MPh every 3 months for at least 2 years, with no change in intervention provided that SF values were maintained below the normal upper limit (males: 300 μg/l; females: 200 μg/l). In this group, SF values ranged from 9 to 300 μg/l, and were, in general, significantly correlated with hepatic iron deposits as assessed by liver R2* (RS = 0·611, P < 0·05). No significant correlation was found between time since the end of IPh (range: 2–28 years) and SF or liver R2* values. The "accepted standard" SF threshold of 99 μg/l (European Association for the Study of the Liver, 2010) identified patients with hepatic R2* higher than 48 Hz (the average R2* value in controls), with 87·5% sensitivity and 85·7% specificity. A SF threshold of 160 μg/l, in turn, identified patients with hepatic R2* higher than 81 Hz (the highest R2* value in controls) with 100% sensitivity and 79·3% specificity, thus supporting this threshold as a safe value to assume hepatic R2* within normal limits (Fig 1A). Importantly, liver steatosis, defined as PDFF >4·8% (França et al, 2017a), was a common finding (67% of MPh patients), particularly when SF was >160 μg/l(80% of these patients) (Fig 1B). This association of liver steatosis with high SF values can be interpreted as steatosis itself giving rise to an increased SF independently of iron load, but it does not rule out the hypothesis that minimal iron toxicity might also be involved in the pathogenesis of steatosis. Our study had limitations. The small number of IPh patients to quantify TBIS hampers the construction of predictive models extensible to larger populations. Although methodological constraints limit iron quantification above a certain load, the iron range observed in this study corresponds to the range of values where more precision is needed in daily clinical practice. A SF threshold for normal liver R2* can be very useful as a practical cut-off, but it does not determine whether iron-induced tissue damage could still be occurring (e.g., as a result of abnormally increased non-transferrin-bound iron). In conclusion, MR determined R2* measurements as surrogate markers of TBIS in patients with HH are feasible. FELIPA, a new variable obtained from liver and pancreas R2* measurements, was correlated with TBIS, calculated by quantitative phlebotomies. FELIPA might be a new useful tool for estimating the severity of iron burden and managing patients' expectations on IPh duration. Liver R2* measurements in HH patients undergoing MPh suggest a best SF cut-off value of 160 μg/l to discriminate between patients with normal iron stores and those in whom there is iron re-accumulation. This study was supported by a research Grant from the Teaching and Research Department of Centro Hospitalar do Porto (grant number DEFI:309/12(213-DEFI/251-CES). The funder had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. Luis Martí-Bonmatí and Ángel Alberich-Bayarri are co-founders of QUIBIM SME. The remaining authors declare that they have no conflicts of interest. Manuela França, Graça Porto, Angel Alberich-Bayarri and Luis Martí-Bonmatí designed the research; Manuela França, Sara Silva, Filipa Vilas Boas, Helena Pessegueiro and Graça Porto performed data acquisition; Manuela França, Graça Porto, Carla Oliveira and Luis Martí-Bonmatí analysed and interpreted the data; Manuela França, Luis Martí-Bonmatí and Graça Porto wrote the paper. Sara Silva, Carla Oliveira, Filipa Vilas Boas, Angel Alberich Bayarri and Helena Pessegueiro Miranda reviewed the paper critically; All the authors approved the submitted version.