Despite the lower rates of cardiovascular morbidity and mortality found in southern Europe,1 cardiovascular disease is the leading cause of morbidity and mortality in Spain.2

Obesity,3 the metabolic syndrome (MS),4 and diabetes mellitus (DM)5 are epidemic in Spain and the industrialized countries, and are associated with morbidity and mortality.6 We are now aware of the importance of adipose tissue in metabolic homeostasis7 and of the fact that its accumulation leads to chronic inflammation and an increase in numerous adipocytokines, such as interleukin 6, monocyte chemotactic protein-1, tumor necrosis factor-α, and leptin, with a decrease in adiponectin. All these changes contribute to insulin resistance and leptin resistance8 and induce the development of dyslipidemia, hypertension, and changes in carbohydrate metabolism, which together constitute MS,9 and, finally, the onset of DM.10 The profiles of inflammatory, metabolic, and other types of biomarkers have not been studied systematically in individuals with obesity, MS, and DM, for comparison with those of healthy individuals.

The main objective of the DARIOS Inflammatory study is to characterize a broad biomarker profile (metabolic, inflammatory, hemodynamic, hemostatic, and myocardial damage) in the metabolic risk continuum defined as the transition from normal weight to obesity, MS, and DM in a Mediterranean population.

The study population consisted of 2851 patients, with a mean age (standard deviation) of 57.4 (8.8) years; 1269 (44.5%) of the participants were men. Table 1 shows the baseline characteristics of the patients in the 4 phenotypic groups, defined by: a) normal weight (with no evidence of MS or DM); b) obesity (with no evidence of MS or DM); c) MS, and d) DM. We observed a progressive and significant trend toward the development of more cardiovascular risk factors, except for the reduction in tobacco use, with an increase in coronary risk, according to the REGICOR equation (the Framingham equation calibrated for the Spanish population), as individuals in the healthy group entered the MS and DM groups. The general characteristics of the DARIOS Inflammatory study and of each of its components are provided in Table S1(supplementary material).

Table 2 shows the values of the different biomarkers analyzed in the 4 groups. The anti-oxidized low-density lipoprotein antibody, interferon-γ, tumor necrosis factor-α, troponin I, and B-type natriuretic peptide titers were not significantly associated with these values. The remaining biomarkers increased significantly, except in the case of adiponectin and apolipoprotein A1, which decreased in individuals who left the healthy group to enter the MS or DM group. Leptin was positively and significantly associated with all 3 disease phenotypes.

Tables S2 and S3 in the supplementary material show the same results when obesity was defined on the basis of waist circumference. In general, the results were similar to those observed when the criterion for obesity was based on BMI.

The correlation between the different biomarkers analyzed was generally weak; the highest correlation coefficient, observed between insulin and leptin, was 0.453 (P > .001) (Table S4 in the supplementary material).

Table 3 shows the results of multinomial logistic regression analysis, taking the group of healthy individuals with normal weight as reference. The phenotype of individuals with normal weight clearly differed from the 3 disease phenotypes studied. Obesity was characterized by a significant positive association with hs-CRP, leptin, and insulin, and a significant negative association with apolipoprotein A1. MS adds to this same pattern a positive association with apolipoprotein B100, which characterizes this condition—single phenotype—and has a significant negative association with adiponectin. DM has a profile similar to that of MS, although its association with all of these same markers, with the exception of hs-CRP, is somewhat weaker.

Upon adjusting the model for HOMA rather than insulin, the results were similar, with the exception of the positive association between leptin and DM, which lost its significance (Table 3). The pattern of associations observed was also similar when obesity was defined on the basis of waist circumference, except in the case of leptin, which maintained the significant positive association with DM (Table S5 in the supplementary material). The study of the interactions demonstrated a significant interaction between hs-CRP and sex, in which increases in hs-CRP were associated with a higher probability of developing obesity, MS, or DM in women than in men.

As this is a multicenter study, the models were also adjusted for the effect of clustering by using random effects models that included the variable “center”, and no statistically significant differences were observed in the associations found.

Finally, Table 4 shows the ability to discriminate between the metabolic risk groups using the biomarkers analyzed here by calculating the area under the curve of the function of Table 3, and Tables S5A and S5B in the supplementary material. In general, the biomarkers exhibited acceptable discrimination between the subgroups and their reference, except for MS vs DM. We also analyzed the calibration of the models (agreement between the probabilities estimated with the models and those actually observed) using the Hosmer-Lemeshow test, and accurate calibration was observed in all the models (except that which included HOMA index and waist circumference to estimate the probability of DM).

In this study, after analyzing a panel of 20 serum biomarkers of carbohydrate metabolism, lipid profile, metabolic status, inflammation, hemostasis, oxidation, hemodynamics, and myocardial damage, we identified a small group of markers (hs-CRP, apolipoprotein A1, apolipoprotein B100, adiponectin, leptin, and insulin resistance) that were independently associated with different phenotypes of metabolic risk, such as obesity, MS, and DM. All of the biomarkers analyzed form part of some of the mechanisms present in the metabolic, inflammatory, and coagulation disorders that are activated (low-grade inflammation) in the different phenotypes studied.

Although the cross-sectional design of the study does not allow us to establish causation, this finding supports the idea that insulin resistance, inflammation, and alterations in the lipid profile play a central role in the continuum of metabolic risk that begins in the healthy individual and progresses to the development of obesity and DM.

Overeating results in adipose tissue remodeling,22 which is characterized by adipocyte hypertrophy, macrophage infiltration, and angiogenesis.23,24 This remodeling favors chronic inflammation,25,26 which can contribute to the development and progression of arteriosclerosis and a certain degree of resistance to insulin action.7,27 This association may not be causal, as occurs in the association between CRP and coronary artery disease,28 but it does indicate the relevance of inflammation in this transition from normal weight to DM. Hyperinsulinemia and insulin resistance (determined using the HOMA index) are common to all 3 phenotypes.

This increased insulin resistance is related to hypoadiponectinemia and hyperleptinemia.29,30 In our study, we also observed that insulin and leptin were the most closely correlated biomarkers, even though the correlation was moderate. It is also interesting that both were independently associated with the phenotypes defined here, which indicates that they represent independent or complementary mechanisms.

The significant positive association between hs-CRP and the phenotypes studied is clear, as is its strong interaction with obesity; the inverse relationship between CRP and adiponectin has been reported elsewhere.31 The causality of this acute phase reactant in atherosclerotic disease is still a subject of debate.32 Chen et al33 reported the presence in blood of several serum proteins that interacted with leptin, and that one of the most important ones was CRP. Thus, they suggested that this protein could contribute to leptin resistance; however, it has not been possible to reproduce these results and they have since been refuted.34

Adiponectin is associated with MS and DM. Adiponectin is a protein that enhances insulin sensitivity and prevents the accumulation of fatty acids in skeletal muscle and internal organs,35 increasing their oxidation. It has also been reported to have anti-inflammatory effects, inhibiting the expression of adhesion molecules, and its possible anti-atherosclerotic effects have also been described.36 Moreover, a number of studies have pointed out that low adiponectin titers are associated with the development of MS, DM, and cardiovascular disease.37–41 The hypoadiponectinemia observed in the majority of these phenotypes is further evidence of the inflammatory state and insulin resistance found in patients with MS or DM.7

Our study demonstrates that this negative association is not yet manifest when obesity is an individual's only diagosed medical condition. This allows us to speculate that normal concentrations of this protein could be useful in the future to identify the phenotype recently described as healthy but obese.42 We have also observed that in the groups with higher cardiovascular risk, estimated according to the REGICOR equation, the adiponectin titers were lower.

Leptin is a protein produced by adipose tissue, with the peculiarity that fat in women produces 3 times more leptin than that of men, and that its function is to prevent fatty acid deposition in nonadipose tissue, restrict food intake, and stimulate growth, fertility, and inflammation.43 In our study, leptin was positively and significantly associated with the 3 phenotypes, although this association lost significance in the diabetic patients when the diagnosis of obesity was based on the BMI and the model was adjusted for the HOMA index. The latter situation was not maintained in the context of obesity diagnosed on the basis of an increased waist circumference.

Apolipoprotein A1 was also inversely associated with all the clinical conditions dealt with here, especially with MS, probably as a consequence of the definition of this syndrome, which includes low high-density lipoprotein cholesterol level as a criterion. Moreover, a decrease in apolipoprotein A1 is also associated with obesity and DM, a circumstance that could be attributed to a more marked insulin resistance. Along these lines, some authors have proposed the use of the triglyceride/high-density lipoprotein cholesterol ratio.44,45

In our study, apolipoprotein B100 is associated only with the presence of MS and, although it is not a criterion for MS per se, it is frequently linked to the lipid abnormalities found in that syndrome. This association indicates that high low-density lipoprotein cholesterol, levels are related to MS, even though it has been observed that the levels in patients with MS may not be very high. However, these particles are small and dense and, consequently, more atherogenic.46 We found no association between the different phenotypes and the biomarkers of lipid oxidation, hemostasis, hemodynamics, or myocardial damage analyzed here. Although plasminogen activator inhibitor-1 was associated with the metabolic phenotypes in the bivariate analysis, this relationship lost statistical significance in the multivariate analysis. This lack of association could be related to the fact that the plasminogen activator inhibitor-1 concentration could not be accurately determined in 23% of the participants, as the concentrations were below the detection limit of the laboratory procedure employed.

Finally, there is a great deal of debate as to the better criterion for the definition of obesity3–47: BMI or waist circumference. In our study, we found no differences between the use of BMI and waist circumference as the criterion for obesity in terms of their association with the biomarkers analyzed here, with the exception of leptin, which lost significance after adjustment of the model for the HOMA index in the context of obesity according to BMI.

Normal weight and, to a lesser extent, obesity differ from MS and DM in their metabolic, inflammatory, and lipid profiles as analyzed by biomarkers, a fact that indicates the relevance of these mechanisms in the continuum of metabolic risk. When MS and DM are compared, these differences are less marked.

This study was fully funded by an unconditional grant from AstraZeneca.

The collection of the original data from the component studies was financed by the ERDF (European Regional Development Fund)-Spanish Ministry of Science and Innovation, Instituto de Salud Carlos III (the HERACLES network RD06/0009); research funds. Spanish Interterritorial Council agreement dated 8 April, 2003; EMER07/046 RCESP C3/09); Fondo de Investigación Sanitaria (Spanish Health Care Research Fund; FIS-ERDF) (PI01/0711, PI02/1158, PI02/1179, PI02/1717, PI03/20471, PI05/2364, PI05/2751, PI07/040, PI07/0934, PI07/1213, G03-045, FIS-ETES 2007, CP06/00100, CM08/00141); Spanish Ministry of Health and Consumer Affairs, National R&D&i Plan 2004-2007, (IP071218); Agència d’Avaluació de Tecnologia i Recerca Mèdica (Agency for the Evaluation of Medical Research and Technology, Catalonia, Spain) (034/33/02); Agència de Gestió d’Ajuts Universitaris i de Recerca (Agency for the Administration of University and Research Grants) (2005SGR00577); Department of Health of the Autonomous Government of Catalonia, Spain; Fundación Canaria de Investigación y Salud (Research and Health Foundation of the Canary Islands, Spain) (45/98); Department of Health of the Government of Navarre; Autonomous Government of Castile and León; Grant for Intensified Research (INT 07/289); Subdirectorate General for the Promotion of Health and Prevention; Department of Health of the Autonomous Community of Madrid; Government of the Balearic Islands; Andalusian Health Service; Program for Community Initiatives of the European Union INTERREG IIIA (SP5.E51); Department of Health of the Autonomous Government of Andalusia, Aid for Research Projects (290/04 y 036/06); Andalusian Society of Family and Community Medicine (SAMFYC 2008); Spanish Society of Family and Community Medicine (semFYC 2009); Department of Health and Consumer Affairs of the Region of Murcia; and Department of Health and Social Welfare, Autonomous Government of Castile-La Mancha.

CIBERESP and CIBEROBEN are an initiative of the Instituto de Salud Carlos III.

None declared.