Ferritin is a 24-subunit evolutionarily conserved protein. Within the cytosol, its main function is to store iron in a soluble nontoxic form, protecting the cell from iron-mediated oxidation-reduction reactions. In the circulation, it acts as a delivery mechanism, and serum levels usually reflect total body iron stores.1 It can be abnormally elevated in a wide range of disease states including malignancy, infection, inflammation, and chronic iron-overload syndromes. Previous investigators have examined the underlying causes of markedly elevated ferritin levels. Liver disease, malignancy, and renal disease have been found to be prominent causes,2 and infections predominated in a recent HIV-positive cohort.3
Rheumatologists are often asked to evaluate patients with markedly elevated serum ferritin levels because of the role it plays as an acute-phase reactant and because of the long-standing association between markedly elevated ferritin levels and rheumatologic diseases such as adult-onset Still’s disease (AOSD)4 or systemic-onset juvenile idiopathic arthritis (SJIA).5 However, making a definitive diagnostic link between a markedly elevated serum ferritin determination and a specific diagnosis, rheumatologic or otherwise, can be challenging as the manifestations of these diseases are often protean and can easily be mistaken for other conditions such as sepsis with multiple organ failure or malignancy.
We reviewed the charts of adult patients with serum ferritin levels greater than 1000 μg/L. We chose this number because it allows for broad clinical applicability in a wide range of patients with hyperferritinemia. These patients were then categorized into specific syndromes known to cause marked hyperferritinemia. We hypothesized that our results would differ from previous studies because of the low prevalence of HIV in our patient population. In addition, over the past several years, there has been an increased appreciation of the association of hemophagocytic lymphohistiocytosis (HLH) with marked hyperferritinemia as well as increased recognition of secondary HLH or macrophage activation syndrome (MAS), in patients with a wide range of primary inflammatory, malignant, and infectious syndromes.6
MATERIALS AND METHODS
Serum ferritin evaluations in our institution are performed using a 2-site immunoenzymatic assay (Beckman Coulter, Brea, CA). Our local clinical laboratory considers a normal serum ferritin to be 30 to 270 μg/L and does not account for patient gender. Our electronic medical record was queried for serum ferritin values greater than or equal to 1000 μg/L occurring from January 1, 2008, through December 31, 2010. The search was limited to patients 18 years or older. For patients having multiple ferritin levels greater than 1000 μg/L, the highest value was used. When available, serum iron and total iron-binding capacity (TIBC) within 30 days of the index ferritin level were recorded, as were erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) if done within 14 days. Each patient’s chart was reviewed, and the most likely cause of hyperferritinemia was decided based on available clinical data. In patients having more than 1 possible etiology, a single predominant underlying cause was decided using a simple algorithm (Fig. 1).
Patients categorized as having iron overload had had monthly red blood cell transfusions for at least 6 months, had been clinically judged to be iron-overloaded by a hematologist, and/or were receiving iron chelation therapy in the setting of a known iron-overload syndrome.
Hepatocellular causes for hyperferritinemia included alcoholic liver disease, active hepatitis B or C infection, cirrhosis (due to nonalcoholic steatohepatitis, primary biliary cirrhosis), autoimmune hepatitis, severe graft-versus-host disease involving primarily the liver, and liver failure due to transient hypotension, toxic ingestion, or graft failure after orthotopic liver transplant. Hyperferritinemia due to severe sepsis or multiple organ dysfunction was classified in this category if hepatic involvement clinically predominated.
Patients were classified as having malignancy if their cancer was active or untreated, or if they were receiving therapy but were within 6 months of their cancer diagnosis. Hematologic malignancies predominated and included myeloma, angioblastic T-cell lymphoma, Hodgkin disease, diffuse large B-cell lymphoma, and mantle cell lymphoma. The most common solid malignancies diagnosed included liver, lung, thyroid, esophageal, breast, renal, pancreatic, colon, and prostate cancers.
Inflammatory causes of hyperferritinemia were composed of those known to be particularly associated with markedly elevated serum ferritin levels (HLH/MAS, Still’s disease) as well as those associated with chronic inflammation. These included flares of systemic lupus erythematosus and flares of granulomatosis with polyangiitis, as well as active rheumatoid arthritis, scleroderma with pericarditis, flare of inflammatory bowel disease, and active graft-versus-host disease. Patients with iron indices consistent with anemia of chronic disease (normocytic anemia and low TIBC) and no other apparent etiology for their elevated ferritin were placed in the separate category of anemia of chronic disease. In addition, patients who were judged by a hematologist to have anemia of chronic disease were also labeled as such.
For the purposes of categorization, possible infectious causes of a markedly elevated ferritin level had to be either acute and severe (ie, requiring inpatient admission and/or aggressive antibiotic therapy) or chronic in nature, requiring long-term antibiotics and/or surgical debridement. Several patients met criteria for SIRS (systemic inflammatory response syndrome), but no infectious source was isolated. These patients were classified in the infectious category based on the treating physician’s clinical judgment that they were, in fact, septic.
Patients placed in the category of chronic kidney disease had to have a documented glomerular filtration rate less than 20 mL/min per 1.73 m2, have stage 4 or 5 chronic kidney disease as assessed by a nephrologist, be on renal replacement therapy, or have chronic kidney disease and be within 3 months of starting renal replacement therapy.
Descriptive statistics were calculated as mean (SD) for continuous variables, and frequency and proportions for categorized values. Wilcoxon rank sum tests were used to compare continuous variables, and Pearson χ2 or likelihood ratio tests were used to compare categorical variables. Statistical analyses were performed using IBM SPSS Statistics, version 20.0 (SPSS Inc, Chicago, IL). Two-sided P ≤ 0.05 was considered statistically significant.
A total of 627 adult patients had a serum ferritin level of at least 1000 μg/L during the study period. Sixty-three percent were white, and 26% were black or African American. There was an overall male preponderance (57.3%). Patient ages ranged from 18 to 91 years (Table 1). The most frequent overall underlying cause of hyperferritinemia was malignancy (24%), with iron-overload syndromes the second most common (22%) (Fig. 1). There were relatively few cases due to purely inflammatory syndromes (n = 27, 4.3%), and of these, only 6 cases due to HLH/MAS or Still’s disease. In 5 cases, no obvious underlying cause could be ascertained. Men often have higher ferritin levels than do women, but when analyzed separately, there were no significant differences between genders in the apparent cause of hyperferritinemia (P > 0.4). In patients younger than 50 years of age, the most common cause was iron overload (27.8%), followed by hepatocellular causes (18%). In patients older than 50 years, malignancy was the dominant cause (30.1%) followed by iron overload (17%) and hepatocellular disease (16.5%) (Table 2).
The mean (SD) serum ferritin level across all categories was 2647 (SD, 3771) μg/L. The highest mean ferritin level was seen in the inflammatory syndromes (4799 [SD, 6830] μg/L), with iron-overload–associated disease the second highest (3112 [SD, 3289] μg/L). The category of anemia of chronic disease had the lowest mean ferritin (1248 [SD, 155] μg/L), whereas those patients with chronic kidney disease had the second lowest (1580 [SD, 1046] μg/L) (Table 3).
Data on serum iron and TIBC were available for 352 subjects, allowing for calculation of transferrin saturation. As expected, the highest mean transferrin saturation was found in the iron-overload category (50% saturation). The lowest transferring saturation was seen in patients with anemia of chronic disease (18.9% saturation) and patients with infections (24% saturation).
There were 18 patients with serum ferritin levels greater than 10,000 μg/L. Within this subgroup, there were 6 patients with a malignancy, 4 with an iron-overload syndrome, 4 with an inflammatory etiology, and 2 each with an infectious or hepatocellular cause.
There were 4 cases of HLH/MAS. Three of these were confirmed by bone marrow biopsy. In the fourth, the diagnosis was made on clinical grounds as bone marrow biopsy was nondiagnostic. Acute infection with Escherichia coli was the triggering factor in 1 case. Another arose in the setting of T-cell lymphoma. The remainder were idiopathic. Cases of AOSD/SJIA were diagnosed clinically using the preliminary criteria of Yamaguchi et al.7 There was 1 case of AOSD (2 major criteria, 3 minor) and 1 adult patient with a history of SJIA (2 major criteria, 2 minor). When analyzed separately, patients with the diagnosis of HLH/MAS/SJIA/AOSD had significantly higher serum ferritin determinations than the overall mean (14,242 [SD, 9379] vs 2535 [SD, 3512], P < 0.001) and in those with other inflammatory conditions (14,242 [SD, 9379] vs 2101 [SD, 2064], P < 0.001) (Fig. 2).
Five patients had no clear diagnosis to explain their hyperferritinemia. A 64-year-old man with a yearlong history of recurrent fever and weight loss was admitted with liver failure, renal failure, and pancytopenia. His ferritin was 5766 μg/L. He underwent an extensive workup, which failed to uncover a unifying diagnosis. He was not evaluated by a rheumatologist. A 30-year-old man presented with recurrent fever, bilateral pleural effusions, and a serum ferritin of 1505 μg/L. He also underwent an extensive evaluation, including rheumatology consultation, without a clear diagnosis being found.
Three patients with no apparent etiology for hyperferritinemia had no clinical illness and appeared well. The mean serum ferritin in this group was 2132 (SD, 912) μg/L. All 3 were previous transplant recipients (2 renal, 1 renal/pancreas) with normal graft function, and all 3 were chronically immunosuppressed. Acute-phase reactants were unavailable for all 3, and iron indices were not consistent with anemia of chronic inflammation.
Serum ferritin level is a nonspecific indicator of systemic illness and can be elevated through a variety of mechanisms, which may overlap or coexist. Systemic inflammation plays a key role in the initiation and perpetuation of hyperferritinemia. Serum ferritin has been widely accepted as an acute-phase reactant and is nonspecifically elevated in a wide variety of inflammatory states including infection, malignancy, and autoimmune diseases. It is particularly elevated in cases of AOSD and HLH, and in fact, the diagnosis of both syndromes rests partially on the presence of a markedly elevated serum ferritin level.7 Diagnostic criteria for HLH include a serum ferritin level greater than 500 μg/L as a minor criterion,8 and an elevated serum ferritin level is seen in approximately 90% of patients with AOSD/SJIA.9 In both of these syndromes, serum ferritin levels can exceed 20,000 μg/L.10 Several groups have established a role for glycosylated ferritin in the diagnosis of inflammatory causes of hyperferritinemia. Fardet et al.11 found significantly lower levels of glycosylated ferritin in patients with hemophagocytic syndrome, and Fautrel et al.12 found that diagnostic criteria for AOSD, which included a glycosylated ferritin fraction of 20% or less, resulted in improved sensitivity and specificity. This assay, however, is as yet not widely used or available, and glycosylated ferritin levels were not available to us in this analysis.
Serum ferritin and systemic iron handling are affected by inflammatory pathways involving the up-regulation of hepcidin. Ferroportin is a cellular membrane protein that allows the egress of iron from enterocytes into the circulation. It is present as well on cells of the reticuloendothelial system, particularly macrophages. Hepcidin binds to and inactivates ferroportin, causing its internalization and degradation within the cell, the net effect of which is to decrease iron absorption from the gut and to limit the amount of iron made available to the circulation, sequestering iron within the reticuloendothelial system.13 Hepcidin production within the liver is stimulated by proinflammatory cytokines, particularly IL-6,14 and it is this constellation of findings (decreased serum iron, increased ferritin, and increased hepcidin levels) that is seen in anemia of chronic inflammation. It is likely that this mechanism at least partially underlies the hyperferritinemia seen in patients with chronic infections, chronic kidney disease, autoimmune disease, and malignancies.
In our patient population, there were very few patients with HLH/MAS or AOSD/SJIA. This may attest to the relative rarity of these diseases, even in a large tertiary medical center. The ferritin cutoff of 1000 μg/L was chosen based on previous work showing that a cutoff of greater than 5 times the upper limit of normal is suggestive of AOSD,15 but it is possible that this value was too high to pick up more of these patients. For example, in a Japanese cohort of patients with AOSD, 6 of 34 patients had normal ferritin levels, and 6 patients with elevated hyperferritinemia had levels less than 5 times the upper limit of normal.16 Although our patients with established Still’s disease or HLH/MAS had markedly elevated serum ferritin levels as one would expect, many more patients with extreme hyperferritinemia had an alternative diagnosis, such as malignancy or infection.
Five of our patients we were unable to clearly categorize. Three of these had undergone solid organ transplantation and were receiving chronic immune suppressive therapy. Although all 3 had normal graft function, it is possible that they all had some level of chronic inflammation, which would account for their elevated serum ferritin level. Iron indices and acute-phase reactants in these patients either were not available or were not indicative of anemia of chronic inflammation. There are data in patients undergoing stem cell transplant for hematologic malignancies showing a poorer prognosis in those with pretransplant iron overload and hyperferritinemia.17 We are not aware of data showing the significance of hyperferritinemia in posttransplant patients with otherwise normal graft function. The 2 remaining patients in this category clearly demonstrated the strong likelihood of an underlying inflammatory process (fever, weight loss, pleural effusions) but, unfortunately, remained undiagnosed.
There are weaknesses in an analysis of this kind. Most prominently, it can be difficult to assign a single diagnosis to a multifactorial outcome. A markedly elevated serum ferritin level can arise from a wide variety of overlapping conditions, and it may be difficult to assign a primary cause in a patient with, for instance, both acute hepatic and chronic renal failure. We attempted to minimize the somewhat arbitrary nature of this assignment by using a simple algorithm, with disease states most associated with hyperferritinemia at the top and those where this association is less profound at the bottom. In this way, we feel we maximized our chances of correctly assigning the most likely etiology causing the highest burden of hyperferritinemia, even in patients with several conditions.
An additional difficulty arises from the nature of retrospective review of patient cases and the limitations of the data available. Up to a quarter of patients with apparent sepsis, for example, do not have documented infection.18 It is possible that there are cases of autoimmune inflammatory disease in our cohort who were incorrectly diagnosed as “culture-negative” sepsis. These patients would not, in general, come to the attention of a rheumatologist, and it is possible that alternative diagnoses may not have been entertained.
Having a higher ferritin cutoff value is typically considered more valuable in predicting the presence of an underlying inflammatory disease. We used a cutoff value greater than 1000 μg/L. We did this because these values are more commonly encountered in the clinical setting, making our analysis more broadly applicable. Although patients in our study with inflammatory conditions such as AOSD or MAS/HLH did have significantly higher serum ferritin levels than those without, the majority of patients with serum ferritin levels greater than 10,000 μg/L carried diagnoses similar to the cohort as a whole.
Finally, our patient population is taken from a tertiary academic medical center and may not reflect a range of diseases or conditions that would more typically be encountered in general practice. In particular, our patients include a significant number of transplant patients, both solid and hematologic. It is possible that the causes of marked hyperferritinemia in a nonacademic medical center would trend more toward malignant and infection causes.
In conclusion, although a markedly elevated serum ferritin level may occur in the presence of an inflammatory disease, it is far more likely to accompany nonrheumatologic conditions such as malignancy, infection, or liver disease. In addition, there are patients with indolent levels of chronic inflammation who may appear well yet have extremely high ferritin levels.
1. Meyron-Holtz EG, Moshe-Belizowski S, Cohen LA. A possible role for secreted ferritin in tissue iron distribution. J Neural Transm
. 2011; 118: 337–347.
2. Lee MH, Means RT. Extremely elevated serum ferritin levels in a university hospital: associated diseases and clinical significance. Am J Med
. 1995; 98: 566–571.
3. Visser A, Mostert C. Causes of hyperferritinaemia classified by HIV status in a tertiary-care setting in South Africa. Epidemiol Infect
. 2012; 15: 1–5.
4. Evensen KJ, Swaak TJG, Nossent JC. Increased ferritin response in adult Still’s disease: specificity and relationship to outcome. Scand J Rheumatol
. 2007; 36: 107–110.
5. Pelkonen P, Swanljung K, Sijmes MA. Ferritinemia as an indicator of systemic disease activity in children with systemic juvenile rheumatoid arthritis. Acta Paediatr
. 1986; 75: 64–68.
6. Allen CE, Yu X, Kozinetz CA, et al. Highly elevated ferritin levels and the diagnosis of hemophagocytic lymphohistiocytosis. Pediatr Blood Cancer
. 2008; 50: 1227–1235.
7. Yamaguchi M, Ohta A, Tsunematsu T, et al. Preliminary criteria for classification of adult Still’s disease. J Rheumatol
. 1992; 19: 424–430.
8. Henter JI, Horne AC, Arico M, et al. HLH-2004: diagnostic and therapeutic guidelines for hemophagocytic lymphohistiocytosis. Pediatr Blood Cancer
. 2007; 48: 124–131.
9. Uppal SS, Al-Mutairi M, Hayat S, et al. Ten years of clinical experience with adult onset Still’s disease: is the outcome improving? Clin Rheumatol
. 2007; 26: 1055–1060.
10. Coffernils M, Soupart A, Pradier O, et al. Hyperferritinemia in adult onset Still’s disease and the hemophagocytic syndrome. J Rheumatol
. 1992; 19: 1425–1427.
11. Fardet L, Coppo P, Kettaneh A, et al. Low glycosylated ferritin, a good marker for the diagnosis of hemophagocytic syndrome. Arthritis Rheum
. 2008; 58: 1521–1527.
12. Fautrel B, Zing E, Golmard J-L, et al. Proposal for a new set of classification criteria for adult-onset Still disease. Medicine
. 2002; 81: 194–200.
13. Nakanishi T, Hasuike Y, Otaki Y, et al. Hepcidin: another culprit for complications in patients with chronic kidney disease? Nephrol Dial Transplant
. 2011; 26: 3092–3100.
14. Nemeth E, Rivera S, Gabayan V, et al. IL-6 mediates hypoferremia of inflammation by inducing the synthesis of the iron regulatory hormone hepcidin. J Clin Invest
. 2004; 113: 1271–1276.
15. Fautrel B, Moel GL, Saint-Marcoux B, et al. Diagnostic value of ferritin and glycosylated ferritin in adult onset Still’s disease. J Rheumatol
. 2001; 28: 322–329.
16. Ohta A, Yamaguchi M, Tsunematsu T, et al. Adult Still’s disease: a multicenter survey of Japanese patients. J Rheumatol
. 1990; 17: 1058–1063.
17. Pullarkat V, Blanchard S, Tegtmeier B, et al. Iron overload adversely affects outcome of allogeneic hematopoietic cell transplantation. Bone Marrow Transplant
. 2008; 42: 799–805.
18. Brun-Buisson C, Doyon F, Carlet J, et al. Incidence, risk factors, and outcome of severe sepsis and septic shock in adults. JAMA
. 1995; 274: 968–974.
© 2013 by Lippincott Williams & Wilkins, Inc.