Metabolic Syndrome Less Strongly Associated With Target Organ Damage Than Syndrome Components in a Healthy, Working Population

Abstract

The authors investigated the associations between target organ damage and individual components of the metabolic syndrome (MS) compared with the MS itself. Carotid intima‐media thickness (IMT), carotid plaque, and left ventricular mass index (LVMI) were assessed by ultrasonography in 356 participants who were free of overt cardiovascular disease. Participants with the MS (n=33) had higher LVMI and carotid IMT than those without the MS (n=323), but the percentage of patients who had carotid plaque was similar. Individually, each component of the MS was significantly associated with the 3 measures of target organ damage. In bivariate and multivariate analyses, the association of clinic systolic blood pressure to both LVMI and carotid IMT and the negative association of high‐density lipoprotein cholesterol with carotid plaque were stronger than and independent of the MS. The data suggest that physicians should evaluate blood pressure and high‐density lipoprotein cholesterol as well as other cardiovascular risk factors without regard to whether a patient meets the criteria for the MS.

The prevalence of the metabolic syndrome (MS) is increasing in developed countries as the prevalence of overweight and obesity increases. ¹ The MS, which represents a cluster of established risk factors and insulin resistance, has been reported to be a strong predictor of coronary heart disease, cardiovascular events, ² , ³ , ⁴ and target organ damage (TOD). ⁵ , ⁶ Although the concept of the MS has been generally accepted and there is little doubt that the risk factors that are used to define it cluster in certain individuals, it is less certain to what extent the adverse outcomes associated with it are due to the combination of multiple risk factors (the MS itself) as opposed to the individual components of the MS. Thus, serious doubts have been raised as to the validity/utility of the MS as a concept, ⁷ which has recently been referred to as a myth. ⁸

Carotid atherosclerosis and left ventricular (LV) hypertrophy are markers of TOD that are associated with both the MS ⁵ , ⁹ and future cardiovascular events. ¹⁰ The relative influence of the different components of the MS (blood pressure [BP], blood lipids, blood glucose, and obesity) on the early development of TOD, however, has not been fully investigated. Thus, the purpose of this study was to compare the relationship of individual MS components, as defined by the National Cholesterol Education Program's Adult Treatment Panel III (NCEP ATP III), ¹¹ with the MS construct itself and measures of subclinical TOD (LV mass index [LVMI]), carotid intima‐media thickness [IMT], and carotid plaque) in healthy working adults without overt cardiovascular disease.

SUBJECTS AND METHODS

Study Population

The study population consisted of 356 participants without evidence of cardiovascular disease who underwent a standard protocol as part of the Work Site Blood Pressure Study, ¹² which was designed to investigate the cross‐sectional and longitudinal relationships of psychosocial factors to BP and TOD. At the time of the evaluation, the average age was 46.9±9.0 years (mean ± SD; range, 30 to 66 years). All participants were employed men and women recruited from 10 work sites in New York City who were normotensive or mildly hypertensive. ¹² Of the 356 participants, 73 (20.5%) were classified as hypertensive according to the criteria of having a clinic BP ≥140/90 mm Hg or taking antihypertensive medication. Following the American Heart Association guidelines, the clinic systolic (and diastolic) BP for each patient was defined as the average of the last 2 (of 3) seated measurements. Hypertensive participants were otherwise healthy; the presence of secondary hypertension was excluded on the basis of clinical and laboratory data. All participants signed an informed consent document and were studied under protocols approved by the institutional review board of Cornell University Medical College.

Echocardiography and Carotid Ultrasonography

Echocardiography and carotid ultrasonography were performed in all participants by an experienced research echocardiography technician according to a standard protocol and using described previous methods. ¹³ M‐mode strip‐chart recordings of the LV were coded and read blindly on as many as 6 high‐quality cycles by a single investigator using a digitizing tablet. Penn convention measurements were used to calculate LV mass. ¹⁴ , ¹⁵ When the M‐mode beam could not be oriented along the LV minor axis from available chest wall acoustic windows, measurements made according to the American Society of Echocardiography recommendations for 2‐dimensional echocardiography ¹⁶ were substituted. LV mass was indexed by estimated body surface area to adjust for differences in body size. LV hypertrophy was defined as LVMI >116 g/m² in men and >104 g/m² in women. ¹⁷

Carotid ultrasonography was performed as described previously. ¹⁸ , ¹⁹ In brief, a 7.0‐ to 7.5‐MHz duplex transducer was used to scan the common, internal, and external carotid arteries for discrete atherosclerotic plaques, defined as the presence of a focal protrusion into the lumen at least 50% greater than the surrounding wall. Individuals with 1 or more discrete plaques were compared with those with no plaques. Two‐dimensional‐guided M‐mode tracings of the distal common carotid artery (approximately 1 cm proximal to the carotid bulb and not at the site of a discrete plaque), with simultaneous electrocardiographic and contralateral carotid pressure waveform tracings, were recorded on videotape and subsequently digitized with a frame grabber and customized software (ARTSS, Cornell Research Foundation, New York, NY). Electronic calipers were used to measure the internal diameter and far‐wall IMT at end‐diastole, recognized from the minimal arterial diameter, as well as the diameter at peak systole. This measurement of the combined IMT of the far wall at end‐diastole has been validated in anatomic correlation studies. ²⁰

Other Measures

Body mass index (BMI) was calculated as body weight (kg) /height² (m²). In addition to clinic BP, 4 components of the MS (waist circumference, high‐density lipoprotein cholesterol [HDL‐C], triglycerides, and fasting glucose) were used for the analysis. Each component of the MS was coded as 1 or 0 (present vs absent) with NCEP ATP III‐defined criteria ¹¹ to calculate metabolic score, and the sum of the dichotomous scores was used as the MS score (0 to 5 points).

STATISTICAL ANALYSES

Group differences in means were compared using analysis of variance. Analysis of covariance and binary logistic regression analyses were used to analyze the relationship between TOD and the MS score (treated as an ordinal categoric scale) while statistically controlling for covariates (Table I). Triglyceride concentration was log‐transformed to reduce its skewness. The associations among continuous measures were evaluated using the Pearson correlation coefficient (Table II) or point‐biserial correlation when one of the measures was binary. Linear regression and logistic regression analyses (Table III) were used to predict the TOD measures and to evaluate the independent predictive utility of the individual components of the MS as well as the MS itself. Change in R ² was used to estimate the independent predictive utility of sets of variables; Nagelkerke's R ² was used similarly in the logistic regression analysis. ²¹ The null hypothesis was rejected when 2‐tailed P<.05. All statistical analyses were performed with SPSS, version 13.0 (SPSS Inc, Chicago, IL).

Table I.

Sample Characteristics by NCEP ATP III Metabolic Syndrome Score

n	Metabolic Syndrome Score				Linear Trend
					P*	P†
	0 (n=148)	1 (n=112)	2 (n=63)	3+ (n=33)	All Categories	Excluding 3+ Category
		Male sex, %	47	56	76‡	76‡	.001	<.001
Race, black, %	22	25	22	15	.367	.923
Age, y	45.3±0.7	46.8±0.8	48.3±1.1	51.6±1.5	.002	.023
Body mass index, kg/m2	24.4±0.3	25.6±0.3	27.8±0.4	30.7±0.6	<.001	<.001
Total cholesterol, mg/dL	206±33	209±47	229±48	219±36	.001	.001
Serum creatinine, mg/dL	0.9±0.2	0.9±0.2	1.0±0.2	0.9±0.1	.012	.003
Factors used for metabolic syndrome
Waist circumference, cm	80±1	85±1	94±1	102±2	<.001	<.001
Fasting glucose, mg/dL	79±9	80±9	85±11	105±54	<.001	<.001
Triglycerides, mg/dL, geometric mean	71±1.5	94±2	145±2	185±2	<.001	<.001
HDL‐C, mg/dL	61±12	51±12	45±13	42±9	<.001	<.001
Clinic SBP, mm Hg	110±1.2	117±1.4	124±1.9	126±2.6	<.001	<.001
Clinic DBP, mm Hg	74±0.7	79±0.8	85±1.1	86±1.5	<.001	<.001
Target organ damage
LVMI, g/m2, unadjusted	75±1.3	77±1.5	85±2.0	86±2.8	<.001	<.001
LVMI, g/m2, adjusted by age	76±1.2	77±1.4	84±1.9	84±2.6	<.001	.001
Carotid IMT, mm, unadjusted	0.68±0.01	0.74±0.02	0.74±0.02	0.82±0.03	<.001	.005
Carotid IMT, mm, adjusted by age	0.69±0.01	0.74±0.01	0.73±0.02	0.78±0.03	.006	.025
Carotid plaque, %, unadjusted	15.5	16.1	34.9‡	24.2	.062	.002
Carotid plaque, %, adjusted by age	14.7	13.3	30.0	16.0	.381	.014
Values are shown as mean ± SE unless otherwise indicated. Each metabolic syndrome component was coded based on National Cholesterol Education Program's Adult Treatment Panel III (NCEP ATP III) criteria. Because the number of subjects with 4 (n=5) or 5 (n=1) metabolic syndrome criteria was so small, they were combined to score 3+. *P shows linear trend. †P shows linear trend when those with metabolic syndrome (3+ category) are excluded. ‡P<.01 vs score 0 by logistic regression analysis. HDL‐C indicates high‐density lipoprotein cholesterol; SBP, systolic blood pressure; DBP, diastolic blood pressure; LVMI, left ventricular mass index; and IMT, intima‐media thickness.

Table II.

Bivariate Correlations of Target Organ Damage Measures With Metabolic Syndrome Components and Metabolic Syndrome Score

	LVMI		Carotid IMT		Carotid Plaque
	R*	P	R*	P	R†	P
Waist circumference, cm	0.381	<.001	0.319	<.001	0.130	.014
Fasting glucose, mg/dL	0.166	.002	0.133	.012	0.065	.225
Log, triglycerides, mg/dL	0.292	<.001	0.203	<.001	0.166	.002
HDL‐C, mg/dL	−0.148	.005	−0.109	.041	−0.129	.015
Clinic SBP, mm Hg	0.404	<.001	0.388	<.001	0.190	<.001
Clinic DBP, mm Hg	0.300	<.001	0.225	<.001	0.136	.010
Metabolic syndrome score	0.163	.002	0.167	.002	0.090	.091
*Pearson's correlation coefficients. †Point‐biserial correlation. LVMI indicates left ventricular mass index; IMT, intima‐media thickness; HDL‐C, high‐density lipoprotein cholesterol; SBP, systolic blood pressure; and DBP, diastolic blood pressure. Metabolic syndrome score is coded from 0 to 5 as shown in Table I.

Table III.

Factors Associated With Target Organ Damage

	Model 1		Model 2		Model 3		Model 4		Model 5
	β	P	β	P	β	P	β	P	β	P
(A) Left ventricular mass index
Sex (male=1, female=0)	0.428	<.001	0.424	<.001	0.318	<.001	0.317	<.001	0.361	<.001
Age, y	0.230	<.001	0.220	<.001	0.148	.004	0.148	.004	0.165	.001
Race (black=1, nonblack=0)	−0.002	.965	−0.001	.991	−0.037	.439	−0.037	.439	−0.036	.446
Metabolic syndrome (yes=1, no=0)			0.068	.133			−0.003	.958
Factors used for metabolic syndrome
Waist circumference, cm					0.060	.289	0.061	.308
Fasting glucose, mg/dL					0.027	.556	0.028	.565
Log triglycerides, mg/dL					0.050	.353	0.051	.358
HDL‐C, mg/dL					0.001	.979	0.001	.981
Clinic SBP, mm Hg					0.190	<.001	0.190	<.001	0.208	<.001
R 2	0.298		0.302		0.337		0.337		0.330
(B) Carotid intima‐media thickness
Sex (male=1, female=0)	0.224	<.001	0.218	<.001	0.147	.011	0.167	.005	0.168	.001
Age, y	0.439	<.001	0.426	<.001	0.396	<.001	0.398	<.001	0.385	<.001
Race (black=1, nonblack=0)	0.006	.902	0.008	.867	−0.016	.741	−0.010	.837	−0.023	.636
Metabolic syndrome (yes=1, no=0)			0.091	.045			0.081	.129
Factors used for metabolic syndrome
Waist circumference, cm					0.048	.405	0.018	.763
Fasting glucose, mg/dL					−0.021	.649	−0.044	.367
Log triglycerides, mg/dL					−0.040	.457	−0.057	.297
HDL‐C, mg/dL					−0.063	.224	−0.059	.255
Clinic SBP, mm Hg					0.167	.001	0.163	.002	0.174	.001
R 2	0.302		0.310		0.331		0.335		0.325
(C) Carotid plaque (logistic regression analysis)
Sex (male=1, female=0)	1.981	.048	2.000	.045	1.576	.260	1.458	.362	1.576	.210
Age, y	1.090	<.001	1.091	<.001	1.095	<.001	1.095	<.001	1.095	<.001
Race (black=1, nonblack=0)	1.824	.105	1.818	.107	1.879	.106	1.833	.122	1.855	.099
Metabolic syndrome (yes=1, no=0)			0.854	.728			0.623	.385
Factors used for metabolic syndrome
Waist circumference, cm					0.987	.379	0.992	.610
Fasting glucose, mg/dL					0.993	.322	0.995	.486
Log triglycerides, mg/dL					1.996	.325	2.292	.248
HDL‐C, mg/dL					0.974	.043	0.973	.037	0.356	.022
Clinic SBP, mm Hg					1.010	.328	1.010	.331
R 2 (Nagelkerke)	0.169		0.169		0.202		0.205		0.191
HDL‐C indicates high‐density lipoprotein cholesterol; and SBP, systolic blood pressure.

RESULTS

In this healthy working population, the prevalence of carotid plaque was 25% in men and 13% in women. In keeping with the mild degree of hypertension in the sample, only 7 of the 356 participants (2%) met the clinical criterion for LV hypertrophy. ¹⁷ There was a higher percentage of men and older participants who met the criteria for the MS, and as expected, they had a higher average BMI (Table IV) than those without the MS. LVMI and carotid IMT in the MS group were significantly higher than those in the non‐MS group, but the percentage of carotid plaque was similar between the groups.

Table IV.

Baseline Characteristics of Subjects Classified by NCEP ATP III‐Defined Metabolic Syndrome

Variables	Metabolic Syndrome		P*
	Absent (n=323)	Present (n=33)
Male sex, %	56	76	.027
Race, black, %	23	15	.383
Age, y	46.4±9.0	51.6±7.3	.002
Body mass index, kg/m2	25.5±3.4	30.7±3.8	<.001
Total cholesterol, mg/dL	211±42	219±36	.337
Serum creatinine, mg/dL	1.0±0.2	0.9±0.1	.665
Factors used for metabolic syndrome
Waist circumference, cm	85±11	102±10	<.001
Fasting glucose, mg/dL	80±10	105±54	<.001
Triglycerides, mg/dL, geometric mean	90±2	185±2	<.001
HDL‐C, mg/dL	54±14	42±9	<.001
Clinic SBP, mm Hg	115±16	126±15	<.001
Clinic DBP, mm Hg	78±9	86±10	<.001
Target organ damage
Left ventricular mass index, g/m2	78±16	86±16	.003
Left ventricular hypertrophy, %	2	3	.497
Carotid intima‐media thickness, mm	0.71±0.17	0.82±0.17	<.001
Carotid plaque, %	19.5	24.2	.498
Values are shown as mean ± SD unless otherwise indicated. *P value testing group differences in percentages and means based on chi‐square test and t test, respectively. NCEP ATP III indicates National Cholesterol Education Program's Adult Treatment Panel III; HDL‐C, high‐density lipoprotein cholesterol; SBP, systolic blood pressure; and DBP, diastolic blood pressure.

To test associations of TOD with the number of MS criteria that were met, we classified the participants into 4 groups based on their MS scores (Table I). All of the TOD measures except for carotid plaque exhibited a dose‐response‐like worsening as the MS scores increased; average LVMI and carotid IMT increased as the MS scores increased. The probability of carotid plaque increased as the MS scores increased from 0 to 2, but did not increase further in patients with an MS score of 3 or higher (Table I). After controlling for age, the linear trend for the relationship of the MS scores remained significant for LVMI and carotid IMT, but not for carotid plaque. When the participants who met criteria for the MS (n=33) were excluded from the analysis, the age‐adjusted linear trend for the MS score was a statistically significant predictor of all 3 measures of TOD.

In bivariate analyses, all 5 components of the MS were significantly correlated with LVMI and carotid IMT, and all but fasting glucose were significantly correlated with carotid plaque (Table II). The MS score was significantly correlated with LVMI and carotid IMT, but the correlations were weaker than with clinic systolic BP, waist circumference, and triglycerides.

With multiple linear regression analyses using the components of the MS as predictors and including race, age, and BMI as covariates, LVMI was positively associated with male sex, age, and clinic systolic BP, but not with other components of the MS or presence of the MS (Table III [A]). When the MS itself and each component of the MS were entered in the same model, only clinic systolic BP was independently associated with LVMI (Table III [A], model 4). Carotid IMT was associated with age, male sex, and clinic systolic BP, but not with the other components (Table III [B]). Although MS itself was significantly associated with IMT when controlling only for covariates (Table III [B], model 2), only clinic systolic BP was significant when the individual components of the MS were included in the model (Table III [B], model 4). Using logistic regression analysis, the presence of carotid plaque was positively associated with age and negatively associated with HDL‐C, but not with any other components of the MS or with the MS itself (Table III [C]). The R ² in models 3 and 5 in Table III [C] were substantially higher than those for models 1 and 2 for each of the 3 outcome measures, indicating that specific individual components of the MS are more closely associated with each measure of TOD than is the MS. The results were essentially unchanged with regard to LVMI and IMT when dichotomous MS components (based on ATP III criteria) were used in the analyses; as expected, the P values for BP became weaker but remained statistically significant, while the P values for the MS remained nonsignificant. The association between dichotomized HDL‐C and carotid plaque became nonsignificant in the same model when using NCEP ATP III criteria (male <40 mg/dL, female <50 mg/dL) to define the cutoff value. When the HDL‐C was dichotomized using the sex‐specific lower quartiles, however, patients in the lowest quartile were significantly more likely to have carotid plaque; the MS was not significantly related to carotid plaque in any model.

DISCUSSION

The main findings of this study are that of the 5 components used to define the MS, clinic BP was independently associated with LV mass and carotid IMT, whereas HDL‐C was independently associated with carotid atherosclerotic plaque even after controlling for other confounders and the presence/absence of the MS. In contrast, although the MS score was positively associated with LVMI and IMT, the presence of the MS itself was a less useful predictor of TOD than were specific individual components (clinic BP or HDL‐C).

In the present study, LVMI was associated with clinic BP independent of other confounders. The relationship persisted when the 33 participants with the MS were excluded from the analysis, and the relationship remained significant when the syndrome itself was added in the multivariate model. LVMI has been reported to be related to age, height, systolic BP, and body size (BMI) in a healthy population, ²² and our results further support this conclusion. In the Strong Heart Study, ²³ only high BP, not the other components of the MS, was associated with LVMI in Native Americans. Although a direct comparison between the individual components of the MS and the MS itself was not performed, that report emphasized that even mildly elevated (nonhypertensive [ie, <140/90 mm Hg]) BP in the MS could increase cardiovascular burden. Schillaci and colleagues ²⁴ also reported that systolic BP was associated with LV hypertrophy independent of the MS in never‐treated hypertensive participants. The present results extend the findings of previous studies to a healthy working population.

Carotid artery IMT is generally considered a valid index of subclinical atherosclerosis ²⁵ and a precursor of overt atherosclerosis. ²⁶ Several studies have found that systolic BP, ²⁷ , ²⁸ , ²⁹ age, ²⁷ , ²⁸ male sex, ³⁰ and hypercholesterolemia ³⁰ , ³¹ are associated with carotid IMT. In the present study, carotid IMT was associated with clinic BP as well as age and male sex, but not with dyslipidemia. The relationships were consistent when the 33 participants with the MS were excluded from the analysis. Although carotid IMT increased as the MS score increased, the relationship of the syndrome to carotid IMT was weaker than that of individual components of the MS, specifically clinic BP. A significant relationship between the MS and carotid IMT has been reported in healthy middle‐aged men, ⁹ healthy volunteers, ³² and elderly women. ³³ Although the latter 2 reports compared the impact of the MS and each component of the MS directly, the characteristics of their populations were very different from those of our population. In one, the age range was wider (21–96 years) and the average age was 10 years higher than in the present sample, ³² while the other studied elderly women whose average age was also much higher than in our population (64 years). ³³ Golden and colleagues ³⁴ examined the relationships between different combinations of the components of the MS and carotid IMT and found that any grouping that included hypertension and triglycerides predicted increased IMT, which indicated that individual components of the MS were more important than the MS overall. That result is consistent with our own, namely that individual components, especially high BP, are more important than the presence of a defined syndrome, the MS, for identifying subclinical carotid atherosclerosis.

In this study, the presence of carotid plaque was negatively associated with HDL‐C independent of the other components of the MS, but was not significantly associated with the MS itself. The age‐adjusted linear trend for carotid plaque with the MS score was significant, however, only when the 33 participants with the MS were excluded (Table I). In the Tromso study, ³⁵ a low level of HDL‐C was associated with an increased risk of having echolucent, rupture‐prone atherosclerotic plaque, which is a strong predictor of stroke. Low HDL‐C has several pathologic associations, including insulin resistance, ³⁶ , ³⁷ impairment of antioxidant and anti‐inflammatory effects, ³⁸ and endothelial dysfunction. ³⁹ There are several papers that have examined the relationship between the MS and carotid artery plaque. Ishizaka and colleagues ⁴⁰ reported that the MS was not associated with carotid plaque in a Japanese population without hypertension. This is in agreement with our results, which indicate that only low HDL‐C, and not the MS, was independently associated with carotid plaque. Our results are also supported by previous studies in which low HDL‐C ⁴¹ and dyslipidemia ⁴² were independently associated with carotid arterial plaque or severity of carotid atherosclerosis independent of other confounders.

It is both a strength and a limitation of our study that the prevalence of the MS in this study (9.3%) was much lower than the age‐adjusted prevalence of 23.7% reported by the third National Health and Nutrition Examination Survey (NHANES III) ⁴³ and that of 27% reported in another work site‐based study in the United States. ⁴⁴ One benefit is that this allowed us to better consider the relationship of the metabolic score to TOD in individuals without overt cardiovascular disease. Because early therapeutic interventions, before development of the MS, are important from a primary prevention perspective, exploring the early correlates of TOD in a middle‐aged healthy population is worthwhile. By comparing the predictive utility of the syndrome with that of the individual MS components, we were able to document the superiority of using individual MS components and the poor predictive value of the MS itself. The coefficients of determination (R ²) in models 3 and 5 are substantially higher than those in models 1 and 2, indicating that the individual components—clinic BP for LVMI and carotid IMT and HDL‐C for carotid plaque—are more strongly associated with TOD.

The disadvantage of this study's sample having a relatively low prevalence of the MS is that we may have underestimated the strength of its association with our multiple measures of TOD. Specifically, in a study in which the prevalence of the MS was greater, we would expect a somewhat higher R ² for model 2 (Table III), and we have little doubt that the statistical significance of the MS would be stronger for this model. We would also expect the R ² for the models with individual MS components to increase, however, because there would necessarily be more patients with “elevated” values of these measures as well. The critical point is that while the low prevalence of the MS may lead us to underestimate the strength of the associations of the MS and its components with TOD, there is no reason to believe that it will bias the coefficients reported in Table III. They clearly demonstrate that for each measure of TOD, a single component of the MS is a better predictor of TOD than is the MS (compare model 5 with model 2) and that a diagnosis of the MS adds little or no predictive value after considering the individual components (model 4).

This study has some additional limitations. The design and analyses were cross‐sectional in nature. Future research should investigate whether these cross‐sectional relationships hold up in prospective analyses. In addition, the fact that the sample consists of healthy, working individuals who volunteered to participate is likely to have resulted in some selection bias associated with a “healthy worker effect”; the predictive utility of the MS might be greater in a less healthy population.

CONCLUSIONS

In a healthy, working population, individual components of the MS (specifically clinic BP and low HDL‐C) were more strongly associated with cardiac and carotid TOD than the MS itself. BP was a strong correlate of LV mass and carotid IMT, and low HDL‐C was most strongly associated with carotid atherosclerotic plaque. These results suggest that primary care physicians should evaluate and treat high BP and low HDL‐C, without regard to whether a patient meets the diagnostic criteria for the MS.

Disclosure: The study was supported in part by NHLBI grants PO1 HL 47540 and R24 HL76857 and by the Banyu Fellowship Program sponsored by Banyu Life Science Foundation International.