Serum Ferritin Is Associated With Visceral Fat Area and Subcutaneous Fat Area

We evaluated the clinical indicators in patients who were admitted to the Yokohama City University Hospital between 2003 and 2005. The protocol was reviewed and approved by the institutional ethical review committee. Informed consent was obtained from all subjects before the examinations. The study was carried out in 248 subjects admitted to our hospital for evaluation of elevated transaminase levels, fatty liver, or stringent control of blood glucose levels. A total of 140 of these were diabetic patients, and 108 were nondiabetic patients. The study was restricted to men and postmenopausal women to eliminate the influence of pregnancy and female hormone replacement therapy and to reduce possible confounding by iron deficiency/ iron supplementation. The study was also restricted to persons >20 years of age to reduce the possible confounding effect of growth and development. A total of 248 Japanese (127 men and 121 postmenopausal women aged 57.8 ± 13.9 years, BMI 25.7 ± 4.6 kg/m², 140 with type 2 diabetes, and 108 without diabetes) subjects were evaluated. Persons with a history of alcohol intake of ≥20 g/day were excluded from the study. Serum ferritin is an acute-phase reactant and does not accurately reflect the body’s iron stores in the presence of inflammation. We therefore measured the serum C-reactive protein in this study to exclude patients with obvious inflammation. The proportion of subjects with history of smoking was 26% in the diabetic group and 13% in the nondiabetic group.

Blood samples were obtained before breakfast on the 2nd day of hospitalization for measurements of the plasma glucose and insulin levels. The insulin resistance was calculated by the homeostasis model assessment of insulin resistance (HOMA-IR) using the following formula: [fasting serum insulin (μU/ml) × fasting plasma glucose (mg/dl)/405]. However, the HOMA-IR and measurement of serum ferritin was performed in only 128 subjects for whom the fasting plasma glucose was <170 mg/dl because HOMA-IR has been reported to be a suitable method for evaluating the presence of insulin resistance in patients with type 2 diabetes only when the fasting glucose levels are <170 mg/dl (9). Ono et al. reported that the results of HOMA-IR were not significantly correlated with the insulin resistance in subjects with fasting glucose levels >200 mg/dl (9).

The abdominal fat distribution in the subjects was determined by computed tomography (CT) with the subjects in the supine position, in accordance with a previously described procedure (10). The SFA and intra-abdominal VFA were measured at the level of the umbilicus using a standardized method in terms of the CT number. In brief, a region of interest was defined in the subcutaneous fat layer by tracing its contour on each scan, and the attenuation range for fat tissue was measured in terms of the CT number (in Hounsfield units).

The degree of liver steatosis was measured by CT. Previous studies have shown a strong correlation between the CT attenuation values of the liver and the extent of fatty infiltration as measured by biopsy (11,12). The ratio of the CT attenuation value of the liver to that of the spleen (L/S ratio) was used for quantitative estimation of the hepatic fat content, with an L/S ratio of <1 being considered to represent fatty liver (11).

All the data were expressed as means ± SD. The relationship between any two variables was analyzed by standard correlation analysis conducted using the StatView version 5.0 software (SAS, Cary, NC). ANOVA was applied for group comparisons, followed by Student’s t test for confirmation of the statistical associations. The relationship between the L/S ratio, VFA, or SFA and other relevant covariates was examined by multiple regression analysis and determination of the standardized correlation coefficients. Statistical significance was assumed when the P value was <0.05.

The baseline characteristics of the 140 patients with type 2 diabetes and 108 nondiabetic patients are shown in Table 1. There were no significant differences in the serum ferritin concentrations, VFA, or SFA between the two groups (type 2 diabetic patients and nondiabetic patients). The baseline characteristics of the 64 diabetic patients for whom the fasting plasma glucose levels were <170 mg/dl and the same number of nondiabetic patients are shown in Table 2. The relationships between the serum ferritin concentrations and the other factors examined in this study are illustrated in Fig. 1. The serum ferritin concentration was inversely correlated with the L/S ratio (r = −0.280, P < 0.0001; Fig. 1A), that is, they were positively correlated with the degree of liver steatosis. The serum ferritin concentration was also significantly correlated with the VFA (r = 0.254, P < 0.0001; Fig. 1B) and SFA (r = 0.231, P = 0.0005; Fig. 1C). In the 64 nondiabetic patients and the 64 patients with type 2 diabetes for whom the fasting plasma glucose levels were <170 mg/dl, the serum ferritin concentration was significantly correlated with the insulin resistance as measured by HOMA-IR (r = 0.286, P = 0.0008; Fig. 1D).

Chronic metabolic and inflammatory diseases may confound the present assessment in diabetic patients. In this study, we excluded subjects with a serum C-reactive protein level of >0.5 mg/dl because serum ferritin is an acute-phase reactant and does not accurately reflect the body iron stores in the presence of inflammation. We also examined the relationship between the serum ferritin concentrations and other factors in the diabetic patients or nondiabetic subjects. The serum ferritin concentration was significantly correlated with the L/S ratio, VFA, SFA, and the degree of insulin resistance in both the diabetic and nondiabetic groups (data not shown). Although Gillum et al. found that the association between the waist-to-hip ratio and the serum ferritin concentration became weaker with increasing age (13), our study revealed no such weakening of the association between the serum ferritin concentrations and the VFA, SFA, and VFA + SFA with age (Table 3).

Multiple regression analysis was conducted to quantify the impact of the measured variables (age, sex, serum ferritin concentration, A1C, and smoking status) on the L/S ratio, VFA, and SFA. The results shown in Table 4 indicate that the age, sex, and serum A1C, but not the serum ferritin concentration, were significantly related to the L/S ratio in all of the study subjects. We then carried out multiple regression analysis separately in the diabetic group and the nondiabetic group. Interestingly, the serum ferritin concentration was found to be significantly correlated with the L/S ratio in the diabetic patients (Table 5), but not in the nondiabetic patients (Table 6). The results shown in Table 7 indicate that the age, sex, serum ferritin concentration, and serum A1C were also significantly correlated with the VFA. The results shown in Table 8 indicate that the sex, serum ferritin concentration, and serum A1C were also significantly correlated with the SFA. The results shown in Table 9 indicate that the serum ferritin and serum A1C were significantly correlated with the HOMA-IR. Smoking status was not an independent variable. Thus, after adjustment for these five variables, the serum ferritin concentration was still found to be correlated significantly with the VFA, SFA, and HOMA-IR, but not the L/S ratio.

In this study, we investigated the association between the serum ferritin concentrations and the hepatic fat content, VFA, SFA, and degree of insulin resistance. We demonstrated that the serum ferritin concentration was significantly associated with the CT-estimated VFA and SFA (Fig. 1B and C). We were also able to reproduce the previously reported association between the serum ferritin concentration and the hepatic fat content and the degree of insulin resistance (4,14) (Fig. 1A and D, Table 4). We also independently analyzed the association between the serum ferritin concentrations and the indexes of distribution of adipose tissues in diabetic patients and nondiabetic subjects. Serum ferritin concentrations were significantly correlated with VFA, SFA, and HOMA-IR in both the diabetic patients and nondiabetic patients (data not shown). Interestingly, there was a significant relationship between serum ferritin concentrations and the L/S ratio in the diabetic patients, but not in the nondiabetic patients (Tables 5 and 6). A previous study on Spanish subjects assessed the associations between obesity and central or abdominal fat distribution and the serum ferritin concentrations (4); they did not find a significant relationship between the serum ferritin and measures of adiposity. Our results, however, were not consistent with these results. This discrepancy may be explained, at least in part, by the reported finding that the waist-to-hip ratio is greater in the Spanish population than in other Caucasian populations (15) and that the ratio is lower in the Japanese population than in western populations.

Gillum et al. reported that the serum ferritin concentration was associated with the waist-to-hip ratio (6). Increased waist-to-hip ratio may suggest an excess VFA in each subject. However, anthropometric measures such as the waist-to-hip ratio do not always distinguish visceral obesity from subcutaneous obesity. To investigate the relationship between the serum ferritin concentration, VFA, and SFA more directly, we elected to use CT in this study and confirmed the relationship between serum ferritin concentrations, VFA, and SFA. After adjustment for age, sex, and serum A1C, serum ferritin concentrations were still significantly correlated with VFA, SFA, and HOMA-IR (Tables 7–9). This is the first report to demonstrate a direct association between the serum ferritin concentrations and VFA and SFA.

Increased iron stores in the liver are postulated to induce liver-mediated insulin resistance, with a reduced ability of insulin to suppress hepatic glucose production (4). Serum ferritin levels have been reported previously to be associated with decreased insulin sensitivity and increased fasting plasma insulin and glucose levels (3,4). These abnormalities might lead to increased adiposity. Several studies have demonstrated correlations between serum ferritin concentrations and other components of the insulin resistance syndrome (3,4,16). It has been reported that the chelation of iron is highly resistant to oxidant-mediated injury (13) and is associated with reduced iron-induced lipid peroxidation (17). It is possible to consider that overexpression of ferritin plays a role as an iron cytoprotective agent by limiting the reactivity of intracellular iron with lipids at the site of their biosynthesis within adipocytes. Thus, ferritin overexpression probably represents adaptive adipocyte response to iron-induced oxidative stress, resulting in a positive association between the serum ferritin concentration and the amount of fat tissue in the body.

Although the hepatic fat content, VFA, and SFA can be evaluated by CT, measurement of the serum ferritin level is a nonradiological and inexpensive test. Therefore, the serum ferritin concentration may be a very useful and cost-effective marker of not only the fat distribution in various organs (hepatic fat content, VFA, and SFA), but also of insulin resistance. Estimation of the serum ferritin concentration may thus be more valuable than CT for evaluation of the fat distribution in subjects with the metabolic syndrome. That is, monitoring of the serum ferritin concentration would be more feasible than serial CT for the follow-up of subjects with the metabolic syndrome. We are now investigating whether changes in the serum ferritin levels after treatment would also be correlated with the changes in the hepatic fat content, VFA, and SFA, as evaluated by CT.

In conclusion, a direct correlation was demonstrated between the serum ferritin concentrations and the VFA and SFA as estimated by CT. Therefore, measurement of the serum ferritin concentration may be a highly useful noninvasive and cost-effective test for the assessment of the fat distribution and degree of insulin resistance.

This work was supported by the Yokohama City University Center of Excellence Program of the Ministry of Education, Culture, Sports, Science and Technology of Japan.