Despite potent insulin-sensitizing, anti-inflammatory, and antiatherogenic effects in animal studies, the relationship between serum adiponectin level and coronary artery disease in patients remains unclear. We determined the adiponectin profile in a cohort of multiethnic Asian patients with coronary artery disease, and the association between serum adiponectin level and culprit lesion necrotic core (NC) content. Ninety-four Asian patients (BMI, 25.3 ± 3.7 kg/m2) undergoing percutaneous coronary intervention were recruited. The serum adiponectin level was measured (n = 94), and the baseline virtual histology intravascular ultrasound examination was analyzed (n = 88). The median level of adiponectin was 3.7 µg/ml (interquartile range, 2.8–4.5 µg/ml). The serum adiponectin level was below 10 µg/ml in 90 patients (95.7%) and below 6 µg/ml in 80 patients (85.1%). There was a significant association between ethnicity and serum adiponectin level (P = 0.048). The median adiponectin level was highest among the Chinese, followed by the Malay and the Indians. Serum adiponectin levels were positively associated with culprit lesion NC content. A 1-µg/ml increase in log adiponectin was associated with a 3.04% (95% confidence interval: 0.33–5.44) increase in culprit lesion NC content. This association remains significant after adjusting for age, sex, ethnicity, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and procedural indication. We found a low serum level of adiponectin in Asian patients and a significant ethnic effect on serum adiponectin level. Increased serum adiponectin levels were independently associated with increased culprit lesion NC burden, suggesting a role for adiponectin in modulating coronary plaque vulnerability.
Adipose tissue is an active endocrine system, and it plays a pivotal role in regulating metabolic and vascular hemostasis through secretion of several biological mediators, collectively known as adipokines. Adiponectin is an adipokine with potent insulin-sensitizing, anti-inflammatory (1,2), and antiatherogenic effects demonstrated in animals (3,4). Yet the relationship between serum adiponectin levels and coronary artery disease (CAD) in patients remains unclear. Several epidemiologic studies have reported decreased expression of adiponectin in patients with CAD and acute coronary syndrome (5,6,7,8). However, subsequent clinical studies in patients with advanced CAD (9,10), heart failure (11), and end-stage renal failure (12,13) have shown that high serum adiponectin levels were independent predictors of mortality and nonfatal cardiovascular events. Further, anecdotal reports on the effect of ethnicity in adiponectin expression have complicated attempts to establish adiponectin as a biomarker (14,15). To date, the role of adiponectin as a clinically useful biomarker for CAD in different populations remains to be defined.
Virtual histology intravascular ultrasound (VH-IVUS) has emerged as an intracoronary imaging technique that allows in vivo evaluation of coronary plaque composition with high accuracy (15,16). Recent studies have demonstrated that the presence of a large necrotic core (NC) on VH-IVUS is associated with acute coronary syndrome and suboptimal results after percutaneous coronary intervention (PCI (17,18)). We sought to profile the adiponectin levels in a cohort of multiethnic Asian patients with advanced CAD undergoing PCI. We further investigated the association between serum adiponectin level and culprit lesion NC content using VH-IVUS.
Methods and Procedures
Study design and patients
The IDEAS (Intravascular ultrasound Diagnostic Evaluation of Atherosclerosis in Singapore)-ADIPO study was a prospective study conducted at the National University Heart Center, Singapore. Asian patients aged 21 years and older who were undergoing PCI for a de novo lesion in a native coronary artery were recruited. Recruited patients would undergo a baseline VH-IVUS examination and blood sampling for adiponectin assay. General exclusion criteria for this study were cardiogenic shock (systolic blood pressure <90 mm Hg), chronic renal failure on dialysis, previous intervention treatment to the target vessel, and an inability to provide informed consent. Angiographic exclusion criteria were angiographically visible residual thrombus despite thrombus aspiration and/or thrombectomy, heavily calcified lesion, tortuous vessel, chronic total occlusion, left main lesion, and distal vessel that was too small to accommodate an IVUS catheter. Demographic and clinical characteristics were collected prospectively by a research nurse. The IDEAS-ADIPO study complies with the Declaration of Helsinki.
The National Healthcare Group Domain Specific Review Board approved the research protocol (reference: C/7/390), and informed consent was obtained from the subjects (or their guardians).
A 6- or 7-French guiding catheter was used to selectively cannulate the ostium of the target coronary artery. A guiding shot was taken after administering a weight-adjusted dose of unfractionated heparin and nitroglycerin (200 µg). All angiographic images captured were analyzed, and offline quantitative coronary analysis (QCA, CAAS, Pie Medical Imaging, Maastricht, the Netherlands) was performed by an experienced angiographer blinded to the results of serum level of adiponectin and VH-IVUS analysis. The minimal lumen diameter, reference segment diameter, percentage of diameter stenosis, and lesion length were measured using standard techniques. Gensini score (19) was used to assess the severity of CAD.
Immediately after guidewire advancement, but before balloon predilation, a 20-MHz, 3.5-French, phased-array IVUS catheter (Eagle-Eye gold; Volcano, Rancho Cordova, CA) was inserted into the target coronary arteries distal to the lesion. Thrombus aspiration was performed if there was angiographic evidence of a thrombus. The IVUS catheter was automatically pulled back to the ostium of the guiding catheter using a motorized pullback device at a speed of 0.5 mm/s (R-100 Pullback device; Volcano). During pullback, gray-scale IVUS images were recorded, and raw radiofrequency ultrasound backscatter data also were captured with electrocardiographic gating at the peak of the R wave for reconstruction of a color-coded map in a VH-IVUS console (S5x Tower; Volcano). All images were recorded in digital format on DVD for subsequent offline quantitative analysis by an investigator (W.K.T.H.) who was blinded to the serum adiponectin level and clinical characteristics of the patients.
Gray-scale IVUS analysis
Conventional gray-scale quantitative IVUS analyses were performed according to the IVUS expert consensus document (20). The lesion site selected for analyses was the image slice with the smallest lumen cross-sectional area. Proximal and distal reference segments were the most normal-appearing cross-sections within 5-mm proximal or distal to the lesion, but before any major side branch. The following parameters were measured at the lesion site: (i) the minimal lumen diameter (mm), (ii) the minimal lumen area (mm2), (iii) the percentage of stenosis (100 × (average reference lumen area−lesion lumen area)/average reference lumen area), (iv) the plaque burden (%), (v) the lesion length, and (vi) the remodeling index (100 × external elastic membrane at lesion/average external elastic membrane area at proximal and distal reference segment).
All target segments, the segment that incorporated the culprit lesion within ±5-mm proximal or distal to the most normal-looking section (atheroma plaque burden <20% by IVUS) was precisely identified and VH-IVUS volumetric analysis was then done. The four VH-IVUS plaque components were color-coded and displayed on the VH-IVUS console: fibrous was shown in green, fibro-fatty in light-green, dense-calcium in white, and NC in red. Volumetric measurements using Simpson's rule over the entire region of interest were performed by tracing the external elastic membrane and lumen border, and the volumetric value of each of the four plaque components in both absolute and relative percentages were then automatically calculated using VIAS 3.0 offline analysis software (Volcano Corp (21)).
Five milli liter of peripheral arterial blood was drawn from the patients' vascular sheaths before heparin administration and the VH-IVUS examination. Blood samples were sent to the department of laboratory medicine for immediately processing and storage at −80°C until analysis.
Serum adiponectin concentrations were measured by Human Adiponectin ELISA Assay (BioVendor, Heidelberg, Germany). Human serum was diluted 1:30 with dilution buffer. Fifty micro liter of diluted standards, samples, and quality controls were incubated in microplate wells precoated with recombinant human adiponectin together with 50 µl of polyclonal antihuman adiponectin antibody conjugated to horseradish peroxidase. The plate was incubated at room temperature for 2 h by shaking at 300 rpm on shaker. After thorough washing, the bound horseradish peroxidase conjugate was allowed to react with 200 µl of substrate at room temperature for 10–15 min. The reaction was stopped by addition of 50 µl acidic solution, and absorbance of the resulting yellow product was measured using a microplate reader set at 450 nm. The product absorbance was inversely proportional to the adiponectin concentration. A standard curve was constructed by plotting absorbance values against adiponectin standards, and concentrations of unknown samples were determined using this standard curve.
Demographic and clinical characteristics of the study subjects involving categorical variables are summarized using frequencies and percentages. For continuous variables, the mean and s.d., or the median and range were used to describe the distribution of the data. As the distribution of adiponectin was skewed, a natural logarithmic transformation was done to normalize the data. The individual effect of demographic and clinical parameters on log adiponectin was evaluated using simple linear regression. The effect of these risk factors was quantified using the beta-coefficient, which corresponds to a 1-µg/ml change in log adiponectin with a unit increase in parameter, together with their associated 95% confidence intervals (CIs). The effect of log adiponectin on each of the IVUS parameters was similarly assessed by simple linear regression. All statistical analyses were generated using STATA software, version 11 (StataCorp LP, College Station, TX), assuming a two-sided test at the conventional 5% level of significance.
Of the 96 eligible patients, 94 were recruited (2 patients declined participation) into the IDEAS-ADIPO study from August 2008 to December 2010. The ethnic composition of the patients was Chinese (n = 61, 65%), Malay (n = 18, 19%), Indian (n = 12, 13%), and other (n = 3, 3%), according to the ethnic composition of Singapore. The mean body height and weight were 1.66 ± 0.07 m and 69.6 ± 11.9 kg, with mean BMI of 25.3 ± 3.7 kg/m2. The demographic and clinical characteristics of the study subjects are shown in Table 1. About 30% of the recruited patients had diabetes mellitus; none had chronic renal failure. The most common indication for the PCI was non-ST-segment elevation myocardial infarction. All the patients were taking aspirin and thienopyridine (clopidogrel, n = 90, prasugrel, n = 4).
Table 1. Demographic and clinical characteristics of study subjects
The left anterior descending artery (72.3%) was the most commonly studied vessel. Over half of the lesions were located at the proximal vessel segment. Offline QCA was performed for all study lesions, and the results are shown in Table 2. The average Gensini score was 44 ± 21.
Table 2. Quantitative coronary and grayscale intravascular ultrasound analysis
The serum level of adiponectin was measured successfully in all 94 recruited patients. The distribution of the serum adiponectin level, which was highly skewed, is shown in Figure 1. The mean and median levels of adiponectin were 4.9 µg/ml and 3.7 µg/ml, respectively. The serum adiponectin level was below 10 µg/ml in 90 patients (95.7%) and below 6 µg/ml in 80 patients (85.1%).
Bivariate relations between log adiponectin and individual demographic or clinical parameters are shown in Table 3. There was a significant association between ethnicity and serum adiponectin level (P = 0.048). The median adiponectin level was lower in the Indians than the Chinese and Malays. In addition, a 1 mmol/l increase in serum high-density lipoprotein cholesterol level was associated with a 0.54 (95% confidence interval: 0.07–1.01, P = 0.025) µg/ml increase in log serum adiponectin level. There was also a tendency of positive association between age and log serum adiponectin level (P = 0.052).
Table 3. Bivariate relationship between log adiponectin and demographic or clinical parameters
Gray-scale and VH-IVUS analyses
Of the 94 patients, IVUS pullback images were available for offline quantitative analysis in 88 patients (93.6%). Among the six patients excluded, the IVUS catheter was unable to cross the lesion in three patients. The IVUS images were suboptimal for analysis in two patients, and not recorded in one patient owing to technical mistakes. There was no incidence of IVUS-related complication. The gray-scale IVUS analysis data are shown in Table 2.
The plaque composition as evaluated by VH-IVUS is shown in Table 4. The median NC content for the minimal lumena area site and entire culprit lesion were both 22.1%. VH-IVUS derived thin-cap fibroatheroma, as defined according to consensus statement, was present in 33 patients (37.5%).
Relation between adiponectin and culprit lesion necrotic core content
The relation between serum adiponectin level, angiographic parameter, and culprit lesion NC content is shown in Table 5. Serum adiponectin level was positively associated with the amount of NC at the entire culprit lesion. In particular, a 1-µg/ml increase in log adiponectin is associated with a 3.04% (95% confidence interval: 0.33–5.44) increase in culprit lesion NC content. Figure 2 illustrates the relationship between NC content and log adiponectin. This association remains significant after adjusting for age, sex, ethnicity, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and PCI indication, as shown in Table 6. When using the mean adiponectin level as a cutoff, the NC content for adiponectin level ≤5 µg/ml (20.7 ± 7.7%) was significantly lower than that for adiponectin level >5 µg/ml (25.0 ± 7.3%, P = 0.030)
Table 5. Bivariate relationship between log adiponectin and intravascular ultrasound/angiographic parameters
Table 6. Multivariable relationship between necrotic core and log adiponectin, adjusted for age, gender, ethnicity, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and procedural indication
Percutaneous coronary intervention and clinical outcomes
All patients underwent successful PCI. Baseline Thrombolysis in Myocardial Infarction (TIMI) flow grade was 0 in 7 (7.4%), 1 in 3 (3.4%), and 3 in 84 patients (89.4%). All patients achieved a final TIMI 3 flow. Median corrected TIMI frame count improved from 28.8 to 17.8. Bare metal stent, drug-eluting stent, and endothelial progenitor cell capturing stent was implanted in 17 (18.1%), 70 (74.5%), and 7 patients (7.4%). PCI-related myocardial infarction occurred in 3 patients (3.2%). At 90-day follow-up, there was one stroke, one heart failure admission, and one target vessel revascularization. There were no deaths or spontaneous myocardial infarctions in our study patients.
We determined the adiponectin profile in 94 multiethnic Asian patients undergoing PCI. Using VH-IVUS technique, we evaluated the association between serum adiponectin level and pre-PCI culprit lesion NC content. We found that the median serum adiponectin level in our cohort of multiethnic Asian patients was lower than that reported in Western populations. The serum adiponectin level was below 10 µg/ml in 96%, and below 6 µg/ml in 85% of the patients. Besides, the serum adiponectin levels were significantly lower in the Indian patients than the Chinese and Malay patients. Serum adiponectin level was positively associated with high-density lipoprotein cholesterol level. For the 88 patients with analyzable VH-IVUS pullback images, we observed a positive correlation between serum adiponectin level and culprit lesion NC content.
There have been limited published data on the effect of ethnicity on serum adiponectin level (14,15). In two studies conducted in the United Kingdom, serum adiponectin level was lower in south Asian than in Western patients. This ethnicity effect was observed in patients with (14) and without CAD (15). The serum adiponectin levels reported in Western populations were over 6–8 µg/ml in most of the studies. For instance, in the AtheroGene study that included 1890 Western patients with CAD (BMI, 27.2 kg/m2), the median serum adiponectin level was 9.12 µg/ml (10). In another study with American patients with CAD (BMI, 30.3 kg/m2), the median serum adiponectin level was 8.5 µg/ml (8). In comparison, the median serum adiponectin level (3.7 µg/ml) in our cohort of multiethnic Asian patients (BMI, 25.3 kg/m2) was much lower. In fact, the serum adiponectin level was below 10 µg/ml in all but four recruited patients. Given the known paradoxical relation between obesity and serum adiponectin level, the unexpectedly low serum adiponectin level in our study patients could be due to the higher percentage of body fat in Asians than Western people of the same age, sex, and BMI (22). Besides, the lack of significant association between serum adiponectin level and BMI in our study suggests that the effect of obesity on adiponectin may be absent in thinner population.
The observation that serum adiponectin level was significantly different among the Chinese, Malay, and Indian patients further supports the effect of ethnicity on adiponectin expression. While the underlying pathophysiological mechanisms remain to be elucidated, this could explain the higher age-standardized event rate of myocardial infarction among the Indians than Chinese and Malay in Singapore (23).
Adiponectin is an endogenous biological mediator secreted mainly by adipocyte tissue. Preclinical studies suggested potent insulin-sensitizing, anti-inflammatory (1,2), and antiatherogenic properties of adiponectin (3,4). Recently, exogenous administration of adiponectin was found to have a protective effect against ischemia-reperfusion injury in an animal model (24).
In humans, however, the data on relations between adiponectin and CAD have been conflicting (5,6,7,8,9,10,11). In particular, studies conducted in patients with advanced CAD, heart failure, and end-stage renal failure have shown a paradoxic relation between serum adiponectin level and clinical outcomes (9,10,11). High serum adiponectin level, instead of being protective, has been shown to be associated with increased mortality and adverse event rates. This paradoxic relation is further supported by our study that included patients undergoing PCI for symptomatic CAD. We demonstrated a positive association between serum adiponectin level and culprit lesion NC content, a plaque phenotype closely associated with acute coronary syndrome.
In contrast to our study, Otake and associates reported a negative association between serum adiponectin level and NC content in 93 Japanese patients presenting with acute coronary syndrome (predominantly unstable angina (17)). The reasons for the discrepancy are not immediately obvious. Both studies included similar numbers of Asian patients, although the ethnic groups studied were different (Japanese vs. Chinese/Malaysian/Indian). Likewise, BMI and cardiovascular risk factor profile of the two study populations are similar. We hypothesize that the discrepancy could be related to different manifestations of the CAD. ST- and non-ST segment elevation myocardial infarction, the most severe form of CAD, was the manifestations in a higher proportion of the patients in our study (53% vs. 18%). This is in keeping with the SAPHIR study, which suggests that the role of adiponectin differs between early atherosclerosis and advanced stage vascular disease (25). Overall, these imaging studies are in accord with the paradoxic relation between adiponectin and CAD at different disease stages (5,6,7,8,9,10,11). Another possible reason for the discrepancy is that both culprit and nonculprit lesions were studied in Otake's study (17). This could have introduced bias, as the relations among adiponectin, nonculprit lesion plaque composition, and clinical manifestation are complex and unclear.
Our finding of positive correlation between adiponectin and culprit lesion NC content is consistent with the hypothesis that in advanced CAD, serum adiponectin levels are elevated in a counter-regulatory fashion (26). High adiponectin level may be a result of excessive oxidative stress and other proatherosclerotic processes, which may explain the association between high serum adiponectin level and high incidence of adverse cardiovascular events.
This is a single-center study and the number of study patients is small. Inherent to the ethnic composition of Singapore, the numbers of Chinese, Malay, and Indian patients in this study were not balanced. Indian and Malay patients were relatively under-represented. Only the level of total adiponectin, but not high molecular weight adiponectin, was measured. Although high molecular weight adiponectin was proposed to be more relevant to CAD (27), contradictory finding exists (11) and high molecular weight adiponectin assay method requires laborious technique for quantification. Arterial rather than venous blood samples were used for adiponectin analysis in our study. Yet, there is no known difference in adiponectin level between arterial and venous blood samples. Owing to the limited resolution and the inability of VH-IVUS to assess fibrous cap thickness, NC content rather than presence of thin-cap fibroatheroma was used as a surrogate for high-risk plaque phenotype. There could be angiographically invisible small thrombi despite thrombus aspiration in patients presenting with ST-segment elevation myocardial infarction, and this could have affected the accuracy of the VH-IVUS analysis. Multiple VH-IVUS endpoints were tested without adjustment for multiple comparisons, and therefore, the possibility of a Type I alpha error cannot be excluded. Levels of other adipokines and biological mediators such as leptin, resistin, high sensitive C reactive protein, and N-terminal pro Brain natriuretic peptide were not measured because of limited funding.
In a cohort of multiethnic Asian patients undergoing PCI, we found that the serum adiponectin level was much lower than that reported in Western populations. In addition, among three ethnic groups, serum adiponectin level was highest in the Chinese, followed by Malay and Indian patients. We also found a positive association between serum adiponectin level and culprit lesion NC content. Future, large-scale studies targeting the effects of different ethnicities on adiponectin expression, as well as the potential of adiponectin as a diagnostic biomarker and therapeutic target, are warranted.
This study was funded by Academic Research Fund (tier 1), Ministry of Education, Singapore (Grant number: R-172-000-181-112). The authors would like to thank the Publication Support Unit, National University Health System for their assistance in the preparation of this manuscript.
W.K.T.H. is a clinical consultant to Valcano Corporation. The other authors declared no conflict of interest.