The global prevalence of obesity is increasing epidemically.

Transcription

1 Circulating Adipocyte Fatty Acid-Binding Protein Levels and Cardiovascular Morbidity and Mortality in Patients With Coronary Heart Disease A 10-year Prospective Study Maximilian von Eynatten, Lutz P. Breitling, Marcel Roos, Marcus Baumann, Dietrich Rothenbacher, Hermann Brenner Objective Adipocyte fatty acid-binding protein (A-FABP) abundantly expressed in mature adipocytes and activated macrophages has dramatic effects on atherosclerosis in mice. Whether this pathophysiological role of A-FABP may also apply to atherosclerotic disease in humans is still unknown. This study investigated associations among serum A-FABP levels, cardiovascular risk factors, and long-term secondary cardiovascular disease (CVD) outcome in patients with coronary heart disease. Methods and Results Serum A-FABP levels were measured in 1069 patients with prevalent coronary heart disease and a 10-year prospective follow-up was conducted (median, [interquartile range, ] months). During this period 204 patients (incidence, 24.0/1000 patient-years) experienced a secondary cardiovascular disease event (defined as cardiovascular death, nonfatal myocardial infarction, or nonfatal cerebrovascular stroke). At baseline, circulating A-FABP was positively associated with a cluster of metabolic and inflammatory risk factors and independently predicted the presence of the metabolic syndrome (odds ratio per unit increase of natural log-transformed A-FABP, 2.95; 95% CI, , P<0.001). On long-term follow-up, subjects with high baseline A-FABP showed an increased risk for secondary cardiovascular disease events (hazard ratio per unit increase, 1.52; 95% CI, ; P=0.001), which was attenuated after multivariable adjustment (hazard ratio 1.30; 95% CI, ). In contrast, A-FABP remained significantly associated with cardiovascular death even after multivariable adjustment (hazard ratio, 1.75; 95% CI, , P=0.007). Conclusion Circulating A-FABP levels are associated with long-term prognosis in patients with coronary heart disease and may represent an important pathophysiological mediator of atherosclerosis, which may point to a new target of treatment options. (Arterioscler Thromb Vasc Biol. 2012;32: ) Key Words: coronary heart disease atherosclerosis secondary prevention metabolic syndrome inflammation lipids epidemiology The global prevalence of obesity is increasing epidemically. 1 A principal mechanistic core of obesity and associated disorders, including the metabolic syndrome (MetS), type 2 diabetes mellitus, and atherosclerosis, resides at the interface of metabolic and inflammatory pathways. 2 Recent insights into the complex mechanisms that link fatty acids and other lipid signals to inflammatory responses have substantially advanced the understanding of atherosclerotic lesion formation. 2,3 Adipocyte fatty acid-binding protein (A-FABP, also designated as ap2 or FABP-4) is a major cytoplasmatic protein that is abundantly expressed in mature adipocytes and activated macrophages. 4,5 It belongs to the superfamily of small molecular weight (14kDa to 15kDa) lipid chaperones known as fatty acid-binding proteins (FABP), a group of molecules that coordinates lipid responses in cells and integrates inflammatory and metabolic responses. 3 In previous animal studies, mice with A-FABP deficiency were protected from developing hyperglycemia, insulin resistance, and dyslipidemia when challenged with genetic and dietary obesity. 6 8 Furthermore, in apolipoprotein E deficient mice, ablation of the A-FABP gene conferred dramatic protection against atherosclerosis in the settings of early and advanced atherosclerosis. 5,9 A recent study demonstrated that an orally active and specific A-FABP inhibitor is highly effective for the treatment of both atherosclerosis and type 2 diabetes mellitus in independent mouse models. 10 Received on: November 28, 2011; final version accepted on: April 24, From the Department of Nephrology, Klinikum rechts der Isar, Technische Universität München, Munich, Germany (M.v.E., M.R., M.B.); Department of Medicine I and Clinical Chemistry, University of Heidelberg, Heidelberg, Germany (M.v.E.); Division of Clinical Epidemiology and Aging Research, German Cancer Research Center, Heidelberg, Germany (L.P.B., H.B.); and Institute of Epidemiology and Medical Biometry, University of Ulm, Ulm, Germany (D.R.). Correspondence to Lutz Breitling, Clinic for Gastroenterology, Hepatology and Infectiology, University Hospital Düsseldorf, Moorenstr. 5, Düsseldorf, Germany. LutzPhilipp.Breitling@med.uni-duesseldorf.de 2012 American Heart Association, Inc. Arterioscler Thromb Vasc Biol is available at DOI: /ATVBAHA

2 2328 Arterioscler Thromb Vasc Biol September 2012 Whether these mechanisms also apply to human disease, where a promoter polymorphism, T-87C, of the A-FABP gene resulting in reduced adipose tissue A-FABP mrna expression has been linked to reduced risk of coronary heart disease (CHD), 11 is not clear yet. Recent studies have demonstrated that A-FABP, although traditionally considered an intracellular cytosolic protein, was released from adipocytes and abundantly present in the circulation, closely correlating with several key features of MetS, including parameters of adiposity, atherogenic dyslipidemia, insulin resistance, hyperglycemia, and hypertension Recently, the association between A-FABP and MetS was confirmed in patients with prevalent coronary artery disease. 15 In an Asian population, A-FABP levels predicted the incidence of MetS and type 2 diabetes mellitus over the course of 5 and 10 years, respectively. 13,14 Furthermore, Yeung et al reported an independent association between circulating A-FABP levels and carotid atherosclerosis. 16 These findings suggest a pathogenetic role of A-FABP in atherosclerotic lesion formation in humans, as has been observed in experimental animals. This is further supported by 2 recent observational longer term follow-up studies in which higher A-FABP was associated with 7-year cardiovascular mortality in 61 patients with end-stage renal disease. 17 Moreover, A-FABP accumulation in atherosclerotic plaques was associated with an unstable plaque phenotype and predicted 3-year adverse cardiovascular events in 561 patients after carotid endarteryectomy. 18 We hypothesized that circulating A-FABP levels are associated with the cluster of metabolic and inflammatory risk factors in patients with manifest atherosclerosis that predispose increased risk for recurrent cardiovascular events among this high-risk population. In addition, we prospectively investigated the 10-year risk of cardiovascular disease (CVD) morbidity and mortality in relation to the baseline A-FABP levels, which would be a readily accessible marker in routine settings, in 1069 patients with prevalent CHD. Methods Study Population Patients with incident CHD (International Classification of Diseases, 9th rev. codes ) aged years and participating in an inhospital rehabilitation program between January 1999 and May 2000 in 2 cooperating clinics (Schwabenland-Klinik, Isny, and Klinik am Südpark, Bad Nauheim, Germany) were enrolled in the study (the total response during baseline recruitment was 58%, resulting in a total number of 1206 patients included in the study at baseline). The design and methods of the study have previously been reported Data Collection and End Point Definition In all patients, active follow-up was conducted 1, 3, 4.5, 6, 8, and 10 years after discharge from the rehabilitation center. Follow-up information and A-FABP measurements were available for 1069 patients (88.6%). Information regarding secondary cardiovascular events and treatment since discharge from the in-hospital rehabilitation clinic was obtained from the primary care physician also by means of a standardized questionnaire. If a patient had died during follow-up, the death certificate was obtained from the local Public Health Department and the main cause of death was coded according to ICD-9 (1-, 3-, or 4.5-year follow-up) or ICD-10 (6-, 8-, or 10-year follow-up). Secondary cardiovascular events were defined either as CVD as the main cause of death (as stated in the death certificate), nonfatal myocardial infarction (MI), or nonfatal cerebrovascular event (stroke). All nonfatal secondary events were reported by the primary care physicians (the last follow-up information was obtained on May 6, 2008). Diagnosis of MetS was based on a modified version of the National Cholesterol Education Program Adult Treatment Panel (NCEP ATP III) and defined as the presence of at least 3 of the following at baseline: body mass index (BMI) >30 kg/m², triglycerides 150 mg/dl ( 1.69 mmol/l), high-density lipoprotein (HDL) cholesterol, men <40 mg/dl (<1.03 mmol/l), women <50 mg/dl (<1.29 mmol/l), blood pressure 130/85 mm Hg, and fasting glucose 110 mg/ dl ( 6.11 mmol/l). In the modified version, BMI instead of waist circumference was used as a measure for adiposity. The study complied with the Declaration of Helsinki and all subjects gave written informed consent. The study was approved by the Ethics Boards of the Universities of Ulm and Heidelberg and of the Physicians chamber of the States of Baden-Wuerttemberg and Hessen (Germany). Clinical Chemistry Blood was drawn at baseline at the end of rehab (mean time from the acute event was 40 days [1st quartile 34 days, 3rd quartile 49 days]) in a fasting state under standardized conditions and stored at 80 C until analysis. Serum A-FABP concentrations were measured by ELISA (BioVendor, Brno, Czech Republic). The intra- and interassay variations were 2.8% and 4.6%, respectively. High sensitive C-reactive protein (hs-crp) and interleukin-6 (IL-6) were determined by a high-sensitivity assay (N Latex CRP mono, Dade Behring, Marburg, and R&D Systems, Wiesbaden, Germany). Serum adiponectin and retinol-binding protein (RBP)-4 levels were measured by ELISA (adiponectin; ALPCO, Salem, NH; RBP-4; Immundiagnostic, Bensheim, Germany). Cystatin C concentrations were measured by immunonephelometry on a Behring Nephelometer II (Dade Behring). Fasting glucose was measured by a glucose oxidase method. Triglyceride, total cholesterol, and HDL cholesterol levels were quantified by standard laboratory methods (all markers measured before rehab discharge) and low-density lipoprotein cholesterol levels were calculated by using the Friedewald formula. All laboratory measurements were performed in blinded fashion. Statistical Analysis The distribution of A-FABP according to sociodemographic characteristics and various cardiovascular risk factors was analyzed by tabulation, age- and sex-adjusted Spearman correlation, and appropriate statistical procedures (χ 2 or Kruskal-Wallis test). In multivariable logistic regression analyses, the independent association of A-FABP levels with the presence of MetS was evaluated. Regression modeling was carried out as available-cases analysis, (ie, excluding observations with missing values only from models in which the respective variable was required). The models included A-FABP, age, and sex (model 1), and additionally adiponectin, RBP-4, and hs-crp levels (model 2), previously shown to be markers of MetS. 22,23 For calculation of the odds ratios, patients were categorized in quintiles based on percentiles of the A-FABP concentrations. The relation of circulating A-FABP levels at baseline (categorized in quintiles) with CVD events during follow-up was assessed by the Kaplan Meier method and tested for significance by means of the log-rank test. We examined the association of A-FABP levels with secondary CVD events corrected for several adjustment sets in multivariable Cox proportional hazards analyses. Model 1: age, sex and BMI; these can be considered classical confounders influencing both A-FABP levels and prognosis and were adjusted for in all models. Model 2: additionally included other traditional cardiovascular risk factors (low-density lipoprotein cholesterol, smoking status, history of hypertension, and creatinine clearance), as well as disease severity indicators associated with prognosis in patients with stable CHD (number of vessels affected, discharge prescription of β-blocking agents, diuretics or angiotensin-converting enzyme inhibitors, and history of MI); to avoid overadjustment, only variables associated with secondary events in age- and sex-adjusted models with P<0.10 or whose inclusion in model 1 changed the A-FABP 5th-quintile

3 von Eynatten et al Serum A-FABP Levels and CVD Risk 2329 coefficient by more than 10% were included at this stage. Adjusting for all variables included in model 2, we further controlled for markers of systemic inflammation (hs-crp and IL-6; model 3), or markers of lipid and glucose metabolism (triglycerides, HDL cholesterol, history of diabetes mellitus, and fasting glucose; model 4), which are assumed to depend on A-FABP and potentially mediate its cardiovascular risk. Significance tests were 2-tailed, and P values <0.05 were considered statistically significant. All statistical procedures were carried out with the SAS statistical software package (release 9.1, 2003, SAS Institute Inc., Cary, NC). The %ROCPLUS macro ( mayoresearch.mayo.edu/mayo/research/biostat/sasmacros.cfm) was used for the calculation of the area under curve and the net reclassification improvement of models predicting outcome events. Results Baseline Characteristics Table 1 shows baseline characteristics with respect to occurrence of secondary CVD events during follow-up. Among the 1069 subjects included in the 10-year follow-up assessment 204 patients experienced a secondary CVD event (91 patients died from CVD, for 60 patients a nonfatal MI and for 53 patients a nonfatal cerebrovascular event was reported). Individuals with an incident secondary CVD event were older and included a larger proportion of patients with diabetes mellitus as compared with patients with event-free survival (Table 1). Moreover, individuals with secondary CVD events had higher levels of A-FABP, triglycerides, fasting glucose, hs-crp, IL-6, systolic blood pressure, and of cystatin C, and a lower creatinine clearance than subjects who did not develop CVD events during follow-up (Table 1). The relationship among various sociodemographic and cardiovascular risk factors and the median of circulating A-FABP levels is summarized in Table 2. The median A-FABP statistically significant increased with older age and was higher in female than in male patients. Presence of MetS and individual components, defined as BMI>30 kg/m², triglycerides 1.69 mmol/l, fasting glucose 6.11 mmol/l, and history of diabetes mellitus and hypertension, were also associated with an increased median of A-FABP. Moreover, A-FABP levels were higher in patients with increased levels Table 1. Baseline Characteristics of Subjects Without and With Secondary CVD Event During 10 Years of Follow-up Characteristics Subjects Without Secondary CVD Event Subjects With Secondary CVD Event P Value* n Age, yr 58.5± ±7.1 <0.001 Men, n (%) 724 (83.7) 180 (88.2) ns BMI, kg/m ± ±3.5 ns Smoking status; n (%) Never smokers 283 (32.7) 52 (25.5) ns Past smokers 541 (62.5) 139 (68.1) Current smokers 41 (4.7) 13 (6.4) Diabetes mellitus, n (%) 129 (15.0) 55 (27.4) <0.001 Hypertension, n (%) 468 (54.3) 121 (60.2) ns Total cholesterol, mmol/l 4.37± ±0.87 ns LDL cholesterol, mmol/l 2.60± ±0.81 ns HDL cholesterol, mmol/l 1.03± ±0.26 ns Trigylcerides, mmol/l 1.41 ( ) 1.53 ( ) Fasting glucose, mmol/l 5.28 ( ) 5.44 ( ) hs-crp. mg/l 3.3 ( ) 4.7 ( ) Interleukin-6, ng/l 3.4 ( ) 4.1 ( ) Adiponectin, mg/l 6.8 ( ) 7.1 ( ) ns RBP-4, µg/l 27.6 ( ) 29.3 ( ) ns Systolic BP, mm Hg 119±15 123± Diastolic BP, mm Hg 73±9 73±9 ns Cystatin C, mg/l 1.02 ( ) 1.14 ( ) <0.001 Creatinine clearance, ml/min 101±28 95±31 <0.001 Lipid-lowering drugs, n (%) 678 (78.6) 147 (72.1) ACE inhibitors, n (%) 442 (51.2) 125 (61.3) Angiotensin receptor blockers, n (%) 30 (3.5) 9 (4.4) 0.52 Metformin, n (%) 41 (4.7) 19 (9.3) Serum A-FABP, µg/l 26.0 ( ) 30.7 ( ) Data are presented as mean±sd or median (interquartile range). CVD indicates cardiovascular death; BMI, body mass index; LDL, low-density lipoprotein; HDL, high-density lipoprotein; hs-crp, high sensitive C-reactive protein; RBP-4, retinol-binding protein-4; BP, blood pressure; ACE, angiotensin-converting enzyme; A-FABP, adipocyte fatty acid-binding protein; ns, not significant. *P value for the difference between subjects with and without secondary CVD events.

4 2330 Arterioscler Thromb Vasc Biol September 2012 Table 2. Serum A-FABP Distribution (Median) According to Various Variables Variable n* Individuals in Top A-FABP Quintile; n (row %) A-FABP Median, µg/l P Value Age, yr (4.3) 17.2 < (11.4) (12.6) 24.0 > (26.2) 30.7 Sex Female (50.3) 48.7 <0.001 Male (14.4) 24.6 BMI kg/m 2 < (12.1) 22.9 < (16.5) 26.5 > (46.1) 43.2 Smoking status Never (23.6) Past (18.8) 25.9 Current 54 6 (11.1) 26.7 Diabetes mellitus No (17.2) 25.6 <0.001 Yes (33.2) 33.6 Hypertension No (13.5) 23.5 <0.001 Yes (25.1) 29.6 Metabolic syndrome No (15.2) 24.9 <0.001 Yes (35.1) 35.6 HDL cholesterol, mmol/l < (19.7) 27.1 ns (19.9) 26.4 Triglycerides, mmol/l < (17.6) 25.4 < (24.1) 30.2 Fasting glucose, mmol/l < (17.4) 25.7 < (29.5) 33.0 hs-crp, mg/l < (8.5) 19.9 < (17.4) 25.2 > (25.2) 30.7 Interleukin-6, ng/l < (10.7) 22.8 < (20.4) 29.1 > (28.9) 31.7 Creatinine clearance, ml/min (18.5) 26.3 <0.001 < (50.0) 47.3 Intake of lipid-lowering drugs No (26.0) 30.6 <0.001 Yes (18.2) 25.8 (Continued) Table 2. Continued Variable n* Individuals in Top A-FABP Quintile; n (row %) A-FABP Median, µg/l P Value Intake of ACE inhibitors No (14.4) 23.4 <0.001 Yes (24.9) 29.7 Intake of angiotensin receptor blockers No (19.8) 26.7 ns Yes 39 9 (23.1) 30.1 Intake of metformin No (19.1) Yes (33.3) 32.7 A-FABP indicates adipocyte fatty acid-binding protein; BMI, body mass index; HDL, high-density lipoprotein; hs-crp, high sensitive C-reactive protein; ACE, angiotensin-converting enzyme. *May not always add up to total because of missing values in few variables. of markers of systemic inflammation, including hs-crp and IL-6 (all P values <0.001). Median A-FABP levels were significantly increased in patients with impaired renal function, with intake of angiotensin-converting enzyme inhibitors or metformin and decreased in subjects receiving lipid-lowering drugs (Table 2). Partial Spearman correlation coefficients between metabolic, inflammatory and other traditional cardiovascular risk factors, and A-FABP are presented in Table 3. In age- and sexadjusted analyses serum A-FABP levels were significantly correlated with BMI, HDL cholesterol, triglycerides, fasting glucose, hs-crp, IL-6, RBP-4, creatinine clearance, cystatin C, and systolic and diastolic blood pressure. For parameters other than creatinine clearance, these associations were not substantially altered when male and female patients were analyzed separately, except for the associations between A-FABP and fasting glucose and diastolic blood pressure that remained clearly significant only in female patients and a significant negative correlation between A-FABP and HDL cholesterol that was found in men only (Table 3). By far the strongest correlation was seen with BMI, and this correlation was particularly strong in women. Serum A-FABP and the MetS On the basis of the NCEP ATP III criteria, MetS was diagnosed in 239 (22.6%) of the 1059 individuals studied (n=10 excluded because of missing values). We performed multivariable logistic regression models to estimate the odds ratios for MetS at baseline across quintiles of serum A-FABP. After sex- and age-adjustment, patients in the second, third, fourth and top compared with the bottom quintile of A-FABP had a significantly increased risk of prevalent MetS (P for all quintiles; Table 4, model 1). Further adjustment for serum adiponectin, RBP-4, and hs-crp levels did not substantively affect this relationship (P for all quintiles; Table 4, model 2). When serum A-FABP was entered as (natural logtransformed) continuous variable in this model, the odds ratio for MetS per unit increase was 3-fold increased: 2.95 (95% CI , P<0.001). To further explore the relationship between A-FABP and MetS in CHD patients, we stratified

5 von Eynatten et al Serum A-FABP Levels and CVD Risk 2331 Table 3. Age- and Sex-adjusted Spearman Correlation Coefficients Between A-FABP Level and Selected Cardiovascular Risk Factors Variable Total Cohort (n=1034)* the median concentrations of serum A-FABP by the number of components of MetS in each individual. The serum concentrations of A-FABP increased significantly as the number of components of MetS increased (0 components: A-FABP median 21.0 [interquartile range, ] µg/l; 1 component: 24.2 [ ] µg/l; 2 components: 28.1 [ ] µg/l; 3 components: 32.4 [( ] µg/l; 4 5 components: 45.9 [ ] µg/l; P<0.001 by Kruskal-Wallis test). A significant association with stepwise increase was also observed when male and female patients were analyzed separately (P<0.001 by Kruskal-Wallis test for both; Figure 1). Serum A-FABP Levels and Risk of Secondary CVD Events After the 10-year follow-up (median follow-up, [interquartile range, ] months) the overall incidence of a secondary CVD event was 24.0 of 1000 patient-years; it was 35.0 per 1000 patient-years in the top quintile of serum A-FABP Men (n=873)* Women (n=161)* r P Value r P Value r BMI < < <0.001 Total cholesterol <0.001 ns ns ns LDL cholesterol ns ns ns HDL cholesterol < ns Triglycerides < < Fasting glucose ns hs-crp < < Interleukin < < <0.001 Adiponectin ns ns ns Retinol-binding protein Systolic BP Diastolic BP < <0.001 Cystatin C < < <0.001 Creatinine clearance ns ns A-FABP indicates adipocyte fatty acid-binding protein; BMI, body mass index; LDL, low-density lipoprotein; HDL, high-density lipoprotein; hs-crp, high sensitive C-reactive protein; BP, blood pressure. *For full comparability of the correlations, only subjects with no missing value in any of the variables are included. P Value distribution compared with 17.3, 19.5, 22.3, and 27.1 per 1000 patient-years in the bottom, second, third, and fourth quintile of the A-FABP distribution (log-rank P=0.006) (Figure 2A). In multivariable Cox proportional regression analyses, subjects with the highest levels of baseline A-FABP had an increased risk of secondary CVD events after adjustment for age, sex, and BMI (hazard ratio [HR] top versus bottom A-FABP quintile, 2.24 [95% CI ]; P=0.002; Table 5, model 1). Controlling for cardiovascular risk factors and disease severity indicators, including low-density lipoprotein cholesterol, smoking status, impaired renal function, discharge prescription of diuretics or angiotensin-converting-enzyme inhibitors, history of MI (HR 1.74 [95% CI ]; P=0.042; Table 5, model 2) and finally, either additional adjustment for markers of systemic inflammation or markers of lipid and glucose metabolism attenuated this relationship (HR 1.66 [95% CI ]; and HR 1.55 [95% CI ]; Table 5, model 3+4). When serum A-FABP was entered as a (natural Table 4. ORs for Prevalent MetS by Quintiles of A-FABP Prevalence MetS (%) A-FABP Median (µg/l) Model 1 OR * (95% CI) P Value Model 2 OR (95% CI) Quintile 1 ( 15.9 µg/l) ref. 1 ref. Quintile 2 (> µg/l) ( ) ( ) Quintile 3 (> µg/l) ( ) < ( ) <0.001 Quintile 4 (> µg/l) ( ) < ( ) <0.001 Quintile 5 (>46.6 µg/l) ( ) < ( ) <0.001 Per unit (continuously) NA NA 2.84 ( ) < ( ) <0.001 ORs indicates odds ratios; MetS, metabolic syndrome; A-FABP, adipocyte fatty acid-binding protein; ref, reference category; NA, not applicable. *Adjusted for age and sex. Additionally adjusted for hs-crp, adiponectin, RBP-4. Natural log-transformed A-FABP. P Value

6 2332 Arterioscler Thromb Vasc Biol September 2012 Figure 1. Serum adipocyte fatty acid-binding protein (A-FABP) concentrations stratified by the number of components of the metabolic syndrome (MetS) analyzed for male and female patients separately (both P<0.001 by Kruskal-Wallis test). The central line represents the median, the boxes span from the 25th to 75th percentiles, and the whiskers extend to the most extreme observations within 1.5 interquartile ranges of the box. Extreme data points are shown individually. log-transformed) continuous variable in the latter models, the HR for secondary CVD events per unit increase was 1.27 (95% CI ) and 1.24 (95% CI ), respectively. Further adjustment for lipid-lowering drugs, or for intake of angiotensin receptor blockers or metformin, had no significant effect on the A-FABP associations with the secondary CVD events. Finally, age- and sex-adjusted HR for subjects with baseline A-FABP in the top quintile did not change substantially after controlling for prevalence of MetS, its individual components, or markers of systemic inflammation (Table 6). Notably, there was a deviation from the proportionality of hazards in these models (ie, an interaction term between logobservation time and the exposure variable was statistically significant [eg, P=0.015 in Table 5, model 2]). This was in line with the Kaplan Meier plot, which suggested a clear separation of the survival curves of the highest 2 quintiles during the earlier follow-up, while there seemed to be more pronounced excess mortality in subjects with lower A-FABP beyond month 96 of follow-up (Figure 2A). In contrast, there was no statistically significant deviation from the proportionality of hazards when analyzing cardiovascular death as the outcome (Figure 2B). The associations of A-FABP with outcome rates appeared more pronounced, with an almost 6-fold risk in the highest versus lowest quintile. Across increasing covariable adjustment sets, the associations were attenuated, but remained statistically significant (Table 5), and the top quintile estimate remained stable when adjusting for MetS variables (Table 6). Additional adjustment for intake of lipid-lowering drugs, angiotensin receptor blockers, or metformin likewise did not affect the results. There was no statistically significant interaction of A-FABP with sex for either outcome (not shown). Evaluating classification performance, the area under the curve of a (logistic) regression model predicting secondary CVD events from all covariables included in model 2 in Table 5 equaled This increased to by inclusion of log-transformed A-FABP. Figure 2. Kaplan Meier estimates for time to secondary cardiovascular disease (CVD) events (A; P=0.006 by log-rank test) and cardiovascular mortality (B; P<0.001 by log-rank test) according to quintiles of adipocyte fatty acid-binding protein (A-FABP). The corresponding net reclassification improvement was (P=0.76). The area under the curve for cardiovascular death increased from to 0.745, with a corresponding net reclassification improvement of (P=0.65). Discussion This 10-year prospective study in 1069 patients with stable CHD at baseline provides the longest observational clinical evidence so far available for serum A-FABP and underlines that this circulating protein may be an important pathoyphysiological mediator of future cardiovascular morbidity and mortality in individuals with manifest atherosclerosis. In this population, increased serum A-FABP levels were also associated with metabolic and inflammatory cardiovascular risk factors at baseline, including adiposity, atherogenic dyslipidemia, hyperglycemia, hypertension, hs-crp, and IL-6. During the long-term follow-up, A-FABP significantly predicted major secondary CVD events (cardiovascular death, nonfatal MI, or ischemic stroke) in crude analyses and for the most impactful outcome of cardiovascular death even after adjustment for established cardiovascular risk factors and lipid-lowering drugs. These findings suggest an independent proatherogenic role of this protein in serum, thereby expanding earlier evidence related to A-FABP as a prognostic cardiovascular risk marker. This was so far based on special populations, comprising end-stage renal disease patients 17 and patients after endarteryectomy in which A-FABP was determined in

7 von Eynatten et al Serum A-FABP Levels and CVD Risk 2333 Table 5. HR and 95% CI for Serum A-FABP at Baseline and Association With Fatal and Nonfatal Cardiovascular Events During Long-term Follow-Up Model 1 Model 2 Model 3 Model 4 HR (95% CI) Adjusted for age, sex and BMI HR (95% CI) *Adjusted for multiple covariates HR (95% CI) Adjusted for multiple covariates HR (95% CI) Adjusted for multiple covariates Fatal and nonfatal Events Serum A-FABP, µg/l 15.9 (bottom quintile) 1 ref. 1 ref. 1 ref. 1 ref. > ( ) 1.05 ( ) 1.01 ( ) 1.02 ( ) > ( ) 1.14 ( ) 1.11 ( ) 1.04 ( ) > ( ) 1.34 ( ) 1.31 ( ) 1.24 ( ) >46.6 (top quintile) 2.24 ( ) 1.74 ( ) 1.66 ( ) 1.55 ( ) P=0.002 P=0.042 Per unit (continuously) 1.52 ( ) 1.30 ( ) 1.27 ( ) 1.24 ( ) P=0.001 Cardiovascular death Serum A-FABP, µg/l 15.9 (bottom quintile) 1 ref. 1 ref. 1 ref. 1 ref. > ( ) P= ( ) 2.03 ( ) 2.39 ( ) > ( ) P=0.017 > ( ) P=0.025 >46.6 (top quintile) 5.98 ( ) P<0.001 Per unit (continuously) 2.29 ( ) P<0.001 atherosclerotic plaques, a tissue not readily available in clinical screening settings. 18 Atherosclerosis is the leading cause of death in developed countries and dysregulation of lipid metabolism and aberrant inflammatory responses represent key features of this disease. 24 As the major storage site for lipids, adipose tissue has been studied extensively with regard to its role in metabolic regulation through lipid signaling. 2 8 However, the exact mechanisms that link lipid signals to inflammatory responses and the formation of atherosclerosis still have to be elucidated. In this context A-FABP, one of the most abundant cytoplasmatic proteins in adipocytes, has been demonstrated to act at the interface of metabolic and inflammatory pathways, thus, exerting a dramatic impact on obesity, type 2 diabetes mellitus, and atherosclerosis in experimental mouse models. 5 10,25 27 Recently A-FABP was shown to be released from adipocytes and abundantly present in human serum. 12 Moreover, circulating A-FABP predicted the development of MetS and type 2 diabetes mellitus in longterm follow-up studies. 13,14 However, the clinical relevance of these findings and its potential pathophysiological implications on atherosclerotic disease in humans is still 2.44 ( ) 2.16 ( ) 2.27 ( ) P= ( ) 1.68 ( ) 1.83 ( ) 3.39 ( ) P= ( ) P= ( ) 3.02 ( ) P= ( ) 1.61 ( ) P=0.023 HR indicates hazard ratio; A-FABP, adipocyte fatty acid-binding protein; BMI. Body mass index; LDL, low-density lipoprotein; ACE, angiotensin-converting enzyme; HDL, high-density lipoprotein. *Adjusted for age, sex, BMI, and covariables selected as described in the methods section (LDL cholesterol, smoking status, impaired renal function, discharge prescription of diuretics, ACE inhibitors, and history of myocardial infarction). Adjusted for above listed variables (*), and high sensitive C-reactive protein and interleukin-6. Adjusted for above listed variables (*), and HDL cholesterol, triglycerides, history of diabetes mellitus, and fasting glucose. Of natural log-transformed A-FABP. unknown. Recent evidence suggests that A-FABP is related to cardiovascular death in patient populations known to be in a proinflammatory state, such as individuals with end-stage renal disease. 17 Our data support this link between serum A-FABP and fatal cardiovascular outcome and it is tempting to speculate that circulating A-FABP could mirror local vulnerability of A-FABP enriched atherosclerotic plaques, 18 a potentially vicious condition that would link A-FABP with fatal atherosclerotic events. Although evidence suggests that adipose tissue is the main source of circulating A-FABP, 13 the proatherogenic activities of this protein appear to be mediated by its direct actions in macrophages. 5 Overexpression of A-FABP in human macrophages directly drives intracellular cholesterol ester accumulation and foam cell formation, 28 whereas A-FABP deficient macrophages exhibit markedly reduced production of proinflammatory cytokines. 29 In our study, we observed a significant positive association between serum A-FABP and markers of systemic inflammation, further supporting the central role of A-FABP in systemic inflammatory responses. Increased systemic inflammation was reported to predict CVD outcome in high-risk patients. 30,31 Findings from this study, however,

8 2334 Arterioscler Thromb Vasc Biol September 2012 Table 6. HRs for Secondary CVD Events Associated With the Top A-FABP Quintile Including Individual Components of the MetS and Systemic Inflammation Variable(s) in Model Fatal and Nonfatal Events suggest that in addition to its proinflammatory trait, A-FABP may influence the atherosclerotic process also through other mechanisms, as a strong and significant association with secondary cardiovascular events persisted even after control for markers of inflammation. Effects of A-FABP on lipid metabolism and associated metabolic responses could act as an alternative pathopyhsiological mediator of atherosclerosis. Lipid disorders have been identified as a key feature in the atherosclerotic process, and A-FABP links fatty acids and other lipid signals to metabolic responses in animals. 7,8 Consistently, serum A-FABP was associated with markers of lipid metabolism in clinical studies, 12,13 and there is considerable evidence that expression, regulation, and metabolic function of A-FABP in humans may be similar to that in mice. In fact, a genetic variant was identified within the promoter region of human A-FABP gene, resulting in genetically determined reduction of A-FABP mrna levels in adipose tissue. 11 Notably, individuals with this genetic variant also had lower triglyceride levels and reduced CVD risk suggesting a pathogenetic role of A-FABP in lipid disorders and associated atherosclerosis, as seen in animals. Our findings that high A-FABP serum levels strongly correlate with triglyceride levels in patients with CHD is consistent with an adverse effect of increased A-FABP expression on lipid metabolism; in fact, this association may partly account for the observed association between circulating A-FABP and secondary CVD events because our results were more influenced by triglycerides than by any other risk factor when included in the statistical models. Indeed, further research is needed to clarify the proposed pathophysiological role of A-FABP in humans and to explore alternative mechanisms by which A-FABP could initiate the atherosclerotic process and deteriorate the progression of this disease. In this context, recent studies directly targeted A-FABP with orally active small-molecule inhibitors and proved this to be an effective therapeutic agent against severe atherosclerosis and type 2 diabetes mellitus in mouse models. 10 Translation of this innovative therapeutic approach into human prospective studies, Cardiovascular Death HR* 95% CI P Value HR 95% CI P Value A-FABP <0.001 MetS +A-FABP <0.001 BMI+A-FABP <0.001 HDL+A-FABP <0.001 TG+A-FABP <0.001 Fasting glucose+a-fabp <0.001 Hypertension+A-FABP <0.001 hs-crp+a-fabp <0.001 Interleukin-6+A-FABP <0.001 HR, hazard ratio; CVD, cardiovascular death; A-FABP, adipocyte fatty acid-binding protein; MetS, metabolic syndrome; BMI, body mass index; HDL, high-density lipoprotein; TG, triglycerides; hs-crp, high sensitive C-reactive protein; NCEP ATP III, National Cholesterol Education Program Adult Treatment Panel. *Bottom A-FABP quintile as reference with HR=1. A-FABP in age- and sex-adjusted models. Diagnosis of MetS according to NCEP ATP III criteria. involving patients at increased risk for vascular disease, will be of great interest. The following limitations of our study should be considered. We could successfully follow-up with 87.2% of the patients. Patients lost to follow-up were, in general, younger and had fewer CVD risk factors (ie, lower BMI and lower CRP values). Although we had a large sample of patients with incident CHD, fatal CVD events were rare in this study population. This may be explained by the fact that mortality of MI is highest during the pre- and early in-hospital phase. In the current study, only patients who were admitted within 3 months after their first acute event had been included. In classification performance analyses, adding A-FABP to well-established cardiovascular risk factors resulted only in minor improvements to the overall predictive value of the models. Thus, our long-term follow-up study does not support a clinical assessment of serum A-FABP for cardiovascular risk stratification, which, however, does not compromise any of the pathophysiological interpretations of the results. Further studies are needed to allow detailed and a priori-specified comparisons of A-FABP to other potential cardiovascular risk markers, such as adiponectin. 21 Interventional studies aiming to reduce serum A-FABP in humans are still lacking. Moreover, extrapolation to current patients should be done with caution, bearing in mind that recruitment of individuals for this study occurred more than 10 years ago. For instance, none of our participants were treated with thiazolidinediones at discharge; usage of these modern drugs could change the associations described in the present study in current patient populations. Finally, outcome assessment may have been imperfect because we had to rely on external information in our study setting including a very large catchment area; this source of error, however, should have been independent from A-FABP and thus should have introduced statistical noise rather than spurious associations in our data. In conclusion, we demonstrate that serum A-FABP levels significantly correlate with inflammatory and metabolic risk factors in patients with prevalent CHD. Moreover, circulating A-FABP is independently associated with a fatal

9 von Eynatten et al Serum A-FABP Levels and CVD Risk 2335 cardiovascular outcome during 10-years follow-up. Our results suggest that A-FABP may represent an important pathophysiological mediator of fatal atherosclerosis in humans and may even point to new treatment targets. The clinical significance of A-FABP as a novel pathophysiological mediator for major atherosclerotic complications should be corroborated in larger long-term follow-up studies, targeting both primary and secondary CVD events. Sources of Funding This cohort study was supported by the German Federal Ministry of Education and Research (Grant 01GD9820/0) and by the Association of German Pension Fund Agencies (Grant ). None. Disclosures References 1. World Health Organization. The World Health Report 2002 Reducing Risk, Promoting Healthy Life. Geneva, Switzerland. World Health Organization Hotamisligil GS. Inflammation and metabolic disorders. Nature. 2006;444: Furuhashi M, Hotamisligil GS. Fatty acid-binding proteins: role in metabolic diseases and potential as drug targets. Nat Rev Drug Discov. 2008;7: Hunt CR, Ro JH, Dobson DE, Min HY, Spiegelman BM. Adipocyte P2 gene: developmental expression and homology of 5 -flanking sequences among fat cell-specific genes. Proc Natl Acad Sci USA. 1986;83: Makowski L, Boord JB, Maeda K, Babaev VR, Uysal KT, Morgan MA, Parker RA, Suttles J, Fazio S, Hotamisligil GS, Linton MF. Lack of macrophage fatty-acid-binding protein ap2 protects mice deficient in apolipoprotein E against atherosclerosis. Nat Med. 2001;7: Hotamisligil GS, Johnson RS, Distel RJ, Ellis R, Papaioannou VE, Spiegelman BM. Uncoupling of obesity from insulin resistance through a targeted mutation in ap2, the adipocyte fatty acid binding protein. Science. 1996;274: Uysal KT, Scheja L, Wiesbrock SM, Bonner-Weir S, Hotamisligil GS. Improved glucose and lipid metabolism in genetically obese mice lacking ap2. Endocrinology. 2000;141: Scheja L, Makowski L, Uysal KT, Wiesbrock SM, Shimshek DR, Meyers DS, Morgan M, Parker RA, Hotamisligil GS. Altered insulin secretion associated with reduced lipolytic efficiency in ap2-/- mice. Diabetes. 1999;48: Boord JB, Maeda K, Makowski L, Babaev VR, Fazio S, Linton MF, Hotamisligil GS. Adipocyte fatty acid-binding protein, ap2, alters late atherosclerotic lesion formation in severe hypercholesterolemia. Arterioscler Thromb Vasc Biol. 2002;22: Furuhashi M, Tuncman G, Görgün CZ, Makowski L, Atsumi G, Vaillancourt E, Kono K, Babaev VR, Fazio S, Linton MF, Sulsky R, Robl JA, Parker RA, Hotamisligil GS. Treatment of diabetes and atherosclerosis by inhibiting fatty-acid-binding protein ap2. Nature. 2007;447: Tuncman G, Erbay E, Hom X, De Vivo I, Campos H, Rimm EB, Hotamisligil GS. A genetic variant at the fatty acid-binding protein ap2 locus reduces the risk for hypertriglyceridemia, type 2 diabetes, and cardiovascular disease. Proc Natl Acad Sci USA. 2006;103: Xu A, Wang Y, Xu JY, Stejskal D, Tam S, Zhang J, Wat NM, Wong WK, Lam KS. Adipocyte fatty acid-binding protein is a plasma biomarker closely associated with obesity and metabolic syndrome. Clin Chem. 2006;52: Xu A, Tso AW, Cheung BM, Wang Y, Wat NM, Fong CH, Yeung DC, Janus ED, Sham PC, Lam KS. Circulating adipocyte-fatty acid binding protein levels predict the development of the metabolic syndrome: a 5-year prospective study. Circulation. 2007;115: Tso AW, Xu A, Sham PC, Wat NM, Wang Y, Fong CH, Cheung BM, Janus ED, Lam KS. Serum adipocyte fatty acid binding protein as a new biomarker predicting the development of type 2 diabetes: a 10-year prospective study in a Chinese cohort. Diabetes Care. 2007;30: Hsu BG, Chen YC, Lee RP, Lee CC, Lee CJ, Wang JH. Fasting serum level of fatty-acid-binding protein 4 positively correlates with metabolic syndrome in patients with coronary artery disease. Circ J. 2010;74: Yeung DC, Xu A, Cheung CW, Wat NM, Yau MH, Fong CH, Chau MT, Lam KS. Serum adipocyte fatty acid-binding protein levels were independently associated with carotid atherosclerosis. Arterioscler Thromb Vasc Biol. 2007;27: Furuhashi M, Ishimura S, Ota H, Hayashi M, Nishitani T, Tanaka M, Yoshida H, Shimamoto K, Hotamisligil GS, Miura T. Serum fatty acidbinding protein 4 is a predictor of cardiovascular events in end-stage renal disease. PLoS ONE. 2011;6:e Peeters W, de Kleijn DP, Vink A, van de Weg S, Schoneveld AH, Sze SK, van der Spek PJ, de Vries JP, Moll FL, Pasterkamp G. Adipocyte fatty acid binding protein in atherosclerotic plaques is associated with local vulnerability and is predictive for the occurrence of adverse cardiovascular events. Eur Heart J. 2011;32: von Eynatten M, Hamann A, Twardella D, Nawroth PP, Brenner H, Rothenbacher D. Relationship of adiponectin with markers of systemic inflammation, atherogenic dyslipidemia, and heart failure in patients with coronary heart disease. Clin Chem. 2006;52: Koenig W, Twardella D, Brenner H, Rothenbacher D. Lipoproteinassociated phospholipase A2 predicts future cardiovascular events in patients with coronary heart disease independently of traditional risk factors, markers of inflammation, renal function, and hemodynamic stress. Arterioscler Thromb Vasc Biol. 2006;26: von Eynatten M, Hamann A, Twardella D, Nawroth PP, Brenner H, Rothenbacher D. Atherogenic dyslipidaemia but not total- and highmolecular weight adiponectin are associated with the prognostic outcome in patients with coronary heart disease. Eur Heart J. 2008;29: Matsushita K, Yatsuya H, Tamakoshi K, Wada K, Otsuka R, Takefuji S, Sugiura K, Kondo T, Murohara T, Toyoshima H. Comparison of circulating adiponectin and proinflammatory markers regarding their association with metabolic syndrome in Japanese men. Arterioscler Thromb Vasc Biol. 2006;26: Qi Q, Yu Z, Ye X, Zhao F, Huang P, Hu FB, Franco OH, Wang J, Li H, Liu Y, Lin X. Elevated retinol-binding protein 4 levels are associated with metabolic syndrome in Chinese people. J Clin Endocrinol Metab. 2007;92: Ross R. Atherosclerosis an inflammatory disease. N Engl J Med. 1999;340: Furuhashi M, Fucho R, Görgün CZ, Tuncman G, Cao H, Hotamisligil GS. Adipocyte/macrophage fatty acid-binding proteins contribute to metabolic deterioration through actions in both macrophages and adipocytes in mice. J Clin Invest. 2008;118: Maeda K, Cao H, Kono K, Gorgun CZ, Furuhashi M, Uysal KT, Cao Q, Atsumi G, Malone H, Krishnan B, Minokoshi Y, Kahn BB, Parker RA, Hotamisligil GS. Adipocyte/macrophage fatty acid binding proteins control integrated metabolic responses in obesity and diabetes. Cell Metab. 2005;1: Boord JB, Maeda K, Makowski L, Babaev VR, Fazio S, Linton MF, Hotamisligil GS. Combined adipocyte-macrophage fatty acid-binding protein deficiency improves metabolism, atherosclerosis, and survival in apolipoprotein E-deficient mice. Circulation. 2004;110: Fu Y, Luo L, Luo N, Garvey WT. Lipid metabolism mediated by adipocyte lipid binding protein (ALBP/aP2) gene expression in human THP-1 macrophages. Atherosclerosis. 2006;188: Makowski L, Brittingham KC, Reynolds JM, Suttles J, Hotamisligil GS. The fatty acid-binding protein, ap2, coordinates macrophage cholesterol trafficking and inflammatory activity. Macrophage expression of ap2 impacts peroxisome proliferator-activated receptor gamma and IkappaB kinase activities. J Biol Chem. 2005;280: Danesh J, Kaptoge S, Mann AG, Sarwar N, Wood A, Angleman SB, Wensley F, Higgins JP, Lennon L, Eiriksdottir G, Rumley A, Whincup PH, Lowe GD, Gudnason V. Long-term interleukin-6 levels and subsequent risk of coronary heart disease: two new prospective studies and a systematic review. PLoS Med. 2008;5:e Sabatine MS, Morrow DA, Jablonski KA, Rice MM, Warnica JW, Domanski MJ, Hsia J, Gersh BJ, Rifai N, Ridker PM, Pfeffer MA, Braunwald E. PEACE Investigators. Prognostic significance of the Centers for Disease Control/American Heart Association high-sensitivity C-reactive protein cut points for cardiovascular and other outcomes in patients with stable coronary artery disease. Circulation. 2007;115:

10 Circulating Adipocyte Fatty Acid-Binding Protein Levels and Cardiovascular Morbidity and Mortality in Patients With Coronary Heart Disease: A 10-year Prospective Study Maximilian von Eynatten, Lutz P. Breitling, Marcel Roos, Marcus Baumann, Dietrich Rothenbacher and Hermann Brenner Arterioscler Thromb Vasc Biol. 2012;32: ; originally published online June 7, 2012; doi: /ATVBAHA Arteriosclerosis, Thrombosis, and Vascular Biology is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 2012 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Arteriosclerosis, Thrombosis, and Vascular Biology can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Arteriosclerosis, Thrombosis, and Vascular Biology is online at: