ADIPOCYTE FATTY ACID BINDING PROTEIN AND C-REACTIVE PROTEIN LEVELS AS INDICATORS OF INSULIN RESISTANCE DEVELOPMENT

Background. Adipocyte fatty acid-binding protein (aFABP) has recently been identified as a potential circulating marker for metabolic syndrome. C-reactive protein (CRP), a sensitive marker of inflammation, is increased in individuals with diabetes and metabolic syndrome due to the development of subclinical inflammation. The study uses logistic regression models to analyze the relations between aFABP and CRP along with other parameters of insulin resistance. The objective was to investigate the potential use of aFABP and CRP levels as tools in the diagnosis of the metabolic syndrome. Methods and results. The following groups were studied: healthy individuals (A, n=122), obese ind1ividuals (B, n=213) and patients diagnosed with metabolic syndrome (C, n=79). Obese persons in Group B had parameters suggestive of early insulin resistance: hypertension, hyperglycaemia, QUICKI (0.305) and higher aFABP levels as compared with the healthy subjects. Group C individuals were diagnosed with metabolic syndrome, as evidenced by the QUICKI markers for insulin resistance (0.293), high aFABP levels (35.3 mg/l). CRP concentrations were lowest in Group A healthy individuals (0.67 mg/l), higher in Group B obese subjects (2.65 mg/l) and highest in Group C patients with metabolic syndrome (3.62 mg/l). Logistic regression models showed an association of aFABP and CRP with BMI (OR 1.12 and 1.39, compared Group A vs C). An association of aFABP and CRP with the QUICKI index showed OR 1.48 and 1.37 (Group A vs C); with triglyceride levels showed OR 1.68 and 1.52 (Group A vs C). An association of aFABP and CRP with glycaemia showed OR 1.48 and 1.51 (Group A vs C), with insulinaemia showed OR 1.44 and 1.38 (Group A vs C) respectively. Conclusions. AFABP levels were higher in obese individuals and highest in those with metabolic syndrome. CRP concentrations were increased in obese persons whereas individuals with metabolic syndrome were found to have high-risk CRP levels. Logistic regression models showed an association of aFABP and CRP with BMI as well as an association of aFABP and CRP with parameters of insulin resistance, namely the QUICKI index, triglyceride levels, glycaemia and insulinaemia. Both methods are of diagnostic benefit for predicting metabolic syndrome, especially in previously untreated patients.


INTRODUCTION
Fat accumulation and adipocyte hypertrophy, accompanying the development of obesity, lead to impaired fat and glucose metabolism and thus the development of insulin resistance.As obesity develops, there is an increase in the discrepancy between adipocyte mass and capacity of the bloodstream in adipose tissue, resulting in adipose tissue hypoxia.In adipose tissue, production of certain adipokines rises, such as that of aFABP whose concentration increases, the development of insulin resistance becomes more pronounced and so is endothelial dysfunction of adipose tissue 1 .AFABP, highly expressed in adipose tissue, has recently been identified as a biomarker for the development of metabolic syndrome based on insulin resistance.The published results suggest that aFABP might be the central regulator of insulin resistance, lipid metabolism and inflammation 2,3 .CRP is a sensitive marker of inflammation.The development of insulin resistance is accompanied by a complex inflammatory reaction with no clinical manifestations.CRP, easily measured in clinical practice, is considered for inclusion in the diagnosis of metabolic syndrome 4 .

PATIENTS AND METHODS
Three groups of subjects were enrolled in the study.Group A comprised 122 healthy individuals (59 males and 63 females) with no signs of insulin resistance.Group B consisted of 213 obese individuals with a BMI ≥ 30 (103 males and 110 females) who were not treated for metabolic disorders.Group C comprised of 79 treated individuals (40 males and 39 females) diagnosed with metabolic syndrome as defined by the NCEP ATP III guidelines 5 .The subjects were enrolled in the study at the Department of Exercise Medicine and Cardiovascular Rehabilitation, University Hospital Olomouc.In all patients, blood pressure, height and weight were measured and blood samples were collected by venipuncture in the morning after 12 h fasting.Basic characteristics of all individuals from the three groups and results of their examination are shown in (Table 1).Height and weight were used to calculate body mass index (BMI).The blood samples were analyzed for glucose, insulin, total cholesterol, triglyceride, HDL cholesterol, CRP, adiponectin and aFABP levels.Serum concentrations of insulin (DPC, USA), glucose, total cholesterol, HDL cholesterol, triglycerides, CRP, adiponectin and aFABP (BioVendor Laboratoty Medicine, Inc., Brno, Czech Republic) were analyzed in fresh serum from morning venous samples taken after 12-hour fasting period.The blood samples were drawn under aseptic conditions from the cubital vein and the serum samples were then obtained in a cooling centrifuge at 3000 g at 4 ˚C and subsequently frozen at -80 ˚C.Concentrations of aFABP and adiponectin were determined after defrosting of individual samples throughout the course of one day.Good Laboratory Practice principles were followed.The study was approved by the Faculty of Medicine and Dentistry, Palacky University Olomouc ethics committee.
The quantitative insulin sensitivity check index (QUICKI) was calculated as follows 6 .QUICKI = 1 / log fasting insulin (μIU/ml) + log fasting glucose (mg/100ml) Statistical analysis was performed using the Statistica 6 software.To test distribution, the Kolmogorov-Smirnov test was used.The statistical significance of differences between values of the control and study groups was assessed by the Mann-Whitney U test.Data in (Table 1) were expressed as means (±standard error SE; Gaussian distribution) or medians (25th and 75th percentiles; non-Gaussian distribution).Spearman's rank correlation coefficient was used to express the relationship of two variables.The significance of CRP and aFABP coefficinents in the logistic models was tested by R-test (www.r-project.org)(ref. 7).

RESULTS
Table 1 shows the results of analyses in all three groups.Variables in the table are presented as the mean or median.All the studied parameters in Groups B and C were different from Group A comprising healthy individuals.The values shown for Group A indicate that all the studied parameters were in normal ranges.Subjects in Group A had high adiponectin levels, low aFABP levels typical for healthy individuals, and very low CRP levels (0.67 mg/l).Obese individuals in Group B had values typical for persons who begin to develop insulin resistance, i.e. hypertension (141/85 mm Hg) and hyperglycaemia (7.0 mmol/l).The QUICKI, which correlates with the clamp techniques for detecting insulin resistance, showed values typical for individuals developing insulin resistance (0.305).As compared with healthy Group A, adiponectin levels were lower (8.6 mg/l) while aFABP levels were higher (29.2 mg/l).CRP levels were slightly increased (2.65 mg/l).Individuals with metabolic syndrome in Group C had levels suggestive of the diagnosis -obesity (BMI 34), hypertension (145/87 mm Hg), hyperglycaemia or type 2 diabetes (8.1 mmol/l) and hypertriglyceridaemia (2.43 mmol/l).The QUICKI levels were typical for individuals with insulin resistance (0.293).aFABP levels were the highest (35.3 mg/l) of all three groups and adiponectin levels were the lowest (7.3 mg/l).CRP concentrations were in the risk range (3.62 mg/l).Table 2 shows statistically significant correlation coefficients between CRP and aFABP and parameters associated with insulin resistance.The highest correlations were between CRP and aFABP (0.46) and between aFABP and BMI (0.57).
Tables 3a and 3b  After adjustment for age, gender and BMI (Model 2), there was a slight association of aFABP and CRP with BMI in both groups (OR 1.06 and 1.25 and 1.12 and 1.39).After adjustment for age, gender and the QUICKI, Model 3 showed a slight association of aFABP and CRP with the QUICKI in both groups (1.13 and 1.32 and 1.48 and 1.37), with the exception of CRP and the QUICKI between the healthy and obese groups (95% CI 0.93-1.37).Model 4, after adjustment for age, gender and triglycerides, showed a slight association of aFABP and CRP with triglycerides in both groups (OR 1.19 and 1.32 and 1.68 and 1.52).Model 5 showed a slight association of aFABP and CRP with glucose in both groups (OR 1.19 and 1.44 and 1.48 and 1.51).Model 6 showed a slight association of aFABP and CRP with insulin in both groups (OR 1.17 and 1.33 and 1.44 and 1.38).

DISCUSSION
The time from the development of insulin resistance to clinical manifestation of components of metabolic syndrome is relatively long (in the order of years).Early detection of potential patients and initiation of intervention are very important for the health of the entire population 6 .The presence of obesity in association with inflammatory reaction accompanied by increased CRP levels is considered a significant predictor of the gradual development of metabolic syndrome.Increased CRP levels are detected in individuals with metabolic syndrome 8 .In Group B obese persons, CRP levels were increased (2.65 mmol/l).The highest CRP levels were in Group C subjects with meta-   9 ) in females with metabolic syndrome were reported by Stefanska 9 .Females should have higher CRP and aFABP levels than males 8 .Logistic regression Model 1 suggested an association between the two parameters and gender.Individuals with fully developed metabolic syndrome may have CRP concentrations of up to 10 mg/l (ref. 10), which was not confirmed by CRP measurements in Group C. Group C subjects have already been treated for metabolic syndrome, most of them with atorvastatin.Statins are known to lower both CRP and aFABP levels 11 .In untreated individuals, the levels of both parameters could be even higher.Therefore, determination of aFABP and CRP levels is useful for early detection of the development of insulin resistance in previously untreated patients.Adipocyte FABP produced by the adipose tissue is considered a potential marker for the development of metabolic syndrome 2 as well as type 2 diabetes 3 .aFABP is considered a predictor of the development of metabolic syndrome based on insulin resistance and overweight or obese individuals are found to have increased levels 12 and so were our groups with obesity and metabolic syndrome.In a group of overweight Korean boys, aFABP levels (23.6 mg/l) (ref. 12) were significantly increased as compared with healthy controls 12 .
Relationship between CRP and aFABP has also been suggested.The logistic regression models assessed the potential use of these markers as tools for estimating the probability of the development of metabolic syndrome.Concentration of aFABP is equivalent with CRP and both have a reasonable diagnostic potential 9 .In accordance with these data, our logistic regression models also showed a slight association of aFABP and CRP with BMI, the QUICKI and triglycerides, related to the development of insulin resistance.Increasing serum aFABP levels are closely associated with impaired glucose metabolism and  may also be considered a useful indicator of the development of type 2 diabetes 13 .Increasing aFABP levels have been shown to be associated with inflammatory markers such as CRP and interleukin-6 in obese individuals and those with metabolic syndrome 14 .The association of AFABP and CRP with metabolic syndrome is often explained by the association with obesity.Some studies reported obesity to be the main factor associating aFABP and CRP (ref. 15,16).Other authors pointed to an association of aFABP and CRP with metabolic syndrome independent of obesity 8,17 .Our analyzed groups adjusted for obesity showed an association of aFABP and CRP with BMI in individuals with both obesity and metabolic syndrome.When assessing aFABP and CRP as diagnostic tools, the sensitivity and specificity of the two methods was evaluated.The results showed similar sensitivity (76%) (ref. 3) and specificity (64%) (ref. 3) levels for both diagnostic methods 3 .Measuring aFABP levels corresponds with the value of CRP in diagnosing metabolic syndrome and assessment of both parameters is of satisfactory (but not excellent) diagnostic potential 9 .

CONCLUSION
Early diagnosis of metabolic syndrome is crucial for decreasing morbidity of populations.Increased aFABP levels accompanied by a rise in CRP levels together with obesity may suggest the development of insulin resistance.Logistic regression models showed an association of aFABP and CRP with BMI as well as an association of aFABP and CRP with parameters of insulin resistance, namely the QUICKI index, triglyceride levels, glycaemia and insulinaemia.Both methods are of diagnostic benefit for predicting metabolic syndrome, especially in previously untreated patients.
compare aFABP and CRP parameters in four logistic regression models.Healthy subjects in Group A are compared with obese Group B individuals in Table 3a and with those with metabolic syndrome (Group C) in Table 3b.Model 1 was adjusted for age and gender, Model 2 for age, gender and BMI, Model 3 for age, gender and the QUICKI, and Model 4 for age, gender and triglycerides, Model 5 for age, gender and glucose, Model 6 for age, gender and insulin.After aFABP and CRP were adjusted for age and gender (Model 1), there was a slight association of aFABP and CRP with gender in both groups (OR 1.19 and 1.34 and 1.47 and 1.43).

Table 1 .
Basic metabolic and clinical characteristics of subjects.

Table 2 .
Statistically significant correlation coefficients among the markers studied (aFABP, CRP) as well as selected parameters of insulin resistance.

Table 3a .
Comparison of CRP and aFABP coefficients significance at healthy subjects in Group A with obese Group B individuals in logistic regression models.

Table 3b .
Comparison of CRP and aFABP coefficients significance at healthy subjects in Group A with metabolic syndrome Group C individuals in logistic regression models.
Model 1 adjusted for sex, age; Model 2 adjusted for sex, age, BMI; Model 3 adjusted for sex, age, Quicki; Model 4 adjusted for sex, age, TAG; Model 5 adjusted for sex, age, glucose; Model 6 adjusted for sex, age, insulin.