Author + information
- Mark D. DeBoer, MD, MSc, MCR∗ (, )
- Matthew J. Gurka, PhD,
- Sherita Hill Golden, MD, MHS,
- Solomon K. Musani, PhD,
- Mario Sims, PhD,
- Abhishek Vishnu, PhD,
- Yi Guo, PhD and
- Thomas A. Pearson, MD, MPH, PhD
- ↵∗Department of Pediatrics, Division of Pediatric Endocrinology, University of Virginia, 409 Lane Road, Room 2017, P.O. Box 800386, Charlottesville, Virginia 22908
There has been controversy whether the metabolic syndrome (MetS) provides additional risk prediction beyond that conferred by levels of the individual MetS components of waist circumference, blood pressure, fasting triglycerides, low high-density lipoprotein, and fasting glucose (1). We previously formulated a continuous MetS severity score (2) that addresses statistical and racial or ethnic limitations of traditional binary MetS criteria such as the Adult Treatment Panel III (ATP-III) criteria. This score was derived from participants of the National Health and Nutrition Examination survey using confirmatory factor analysis of the 5 MetS components on a sex- and race or ethnicity–specific basis (2). This approach provided differential weights (the largest being for waist circumference and the smallest being for systolic blood pressure) that are multiplied by a given individual’s absolute values for each MetS component toward a continuous z-score, estimating the degree of MetS severity by sex and racial or ethnic subgroup, with a mean of 0 and increasing values indicating higher MetS severity. This MetS severity score enables following MetS severity over time and has been shown in a small sample to correlate with future cardiovascular disease and diabetes (3).
In the current study, we utilized adjudicated coronary heart disease (CHD) outcomes from the ARIC (Atherosclerosis Risk in Communities) study (11,004 white and black participants recruited from 1987 to 1989 with mean age 53.8 years and ≤24 years’ follow-up) and JHS (Jackson Heart Study) study (2,137 black participants recruited from 2000 to 2004 at mean age 48.5 years with ≤11 years’ follow-up) (1,4). We assessed abilities of baseline MetS (both ATP-III MetS classification and MetS severity z-scores) (1,2) on a sex- and race-specific basis to predict future CHD in Cox proportional hazards models (adjusted for age and stratified by site) both without and with the 5 MetS components. This cause-specific analysis (with non-CHD death censored) was compared to a formal competing risk analysis using the Fine and Gray model. The ability of MetS severity to discriminate future CHD was calculated by the area under the curve of a time-dependent receiver-operating characteristic curve, and compared with ATP-III MetS classification. The ability of MetS severity to improve classification of future disease beyond that of ATP-III MetS classification was calculated by the continuous net reclassification improvement (NRI) and integrated discrimination improvement (IDI) after fitting both MetS metrics in a Cox model. In addition to the continuous NRI, we also calculated the event NRI (which can be interpreted as the net percentage of participants with a CHD event who were correctly assigned a higher predicted risk when adding MetS severity) as well as the 10-event NRI (the net percentage of participants without CHD correctly assigned a lower risk) (5).
From the competing-risks analysis, ATP-III MetS classification was significantly associated with future CHD (hazard ratio [HR]: 1.64; 95% confidence interval [CI]: 1.49 to 1.80; p < 0.01). Notably, we observed a significant interaction between ATP-III MetS classification and sex (male HR: 1.37; female HR: 2.09), and following adjustment for the individual MetS components, ATP-III MetS classification was no longer associated with future CHD (HR: 0.87; 95% CI: 0.73 to 1.05; p = 0.10). The MetS severity z-score was associated with future CHD in models both without and with the individual MetS components (both p <0.01) (adjusted HRs shown in Figure 1). Figure 1 displays model-estimated cumulative incidence functions for the 4 categories of MetS severity. Participants in the fourth quartile of MetS severity had a 25-year CHD rate of 24.9% compared to a 6.5% rate for those in the first quartile of MetS severity. There were no significant sex or race interactions, nor was there excessive collinearity of the score with the individual components (variance inflation factors <10). Cause-specific Cox models provided similar HRs, and were used as the basis for model-prediction reclassification assessment. The time-dependent area under the curve was 0.56 for ATP-III MetS classification and 0.63 for the MetS z-score, supporting improved predictive ability. Both the NRI and the IDI, measures of risk reclassification and discrimination improvement, respectively, were significantly >0 when adding MetS z-score to a survival model of CHD that originally included ATP-III MetS classification (again without excess collinearity): continuous NRI was 0.30 (95% CI: 0.24 to 0.36; p < 0.001); event NRI was 0.13 (95% CI: 0.07 to 0.19); the nonevent NRI was 0.17 (95% CI: 0.15 to 0.19); and the IDI was 0.017 (95% CI: 0.016 to 0.018; p < 0.001).
Although MetS has been criticized as not being worth more than the sum of its individual parts, these data demonstrate that a continuous estimate of MetS severity provided additional predictive ability for future CHD beyond the individual MetS components. This score may thus contribute additional risk related to the underlying pathophysiological processes of cellular dysfunction, oxidative stress, and systemic inflammation that appear to produce abnormalities in the individual MetS components. Moreover, this sex- and race-specific score, unlike ATP-III MetS classification, did not exhibit differences in CHD prediction by sex or race and thus may provide similar risk information across sub-populations. Overall, the MetS severity z-score performed better than and appeared to offer multiple advantages over ATP-III MetS classification in CHD risk prediction. Cutoff values indicating particular increase in MetS severity–related risk are still needed.
Please note: This work was supported by grants from the National Institutes of Health (Bethesda, Maryland). 1R01HL120960 (Drs. DeBoer and Gurka). The Jackson Heart Study is supported by contracts HHSN268201300046C, HHSN268201300047C, HHSN268201300048C, HHSN268201300049C, and HHSN268201300050C from the National Heart, Lung, and Blood Institute and the National Institute on Minority Health and Health Disparities. The Atherosclerosis Risk in Communities Study is carried out as a collaborative study supported by National Heart, Lung, and Blood Institute contracts (HHSN268201100005C, HHSN268201100006C, HHSN268201100007C, HHSN268201100008C, HHSN268201100009C, HHSN268201100010C, HHSN268201100011C, and HHSN268201100012C). The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. DeBoer and Gurka contributed equally to this work.
- American College of Cardiology Foundation
- McNeill A.M.,
- Schmidt M.I.,
- Rosamond W.D.,
- et al.
- DeBoer M.D.,
- Gurka M.J.,
- Woo J.G.,
- Morrison J.A.
- Xanthakis V.,
- Sung J.H.,
- Samdarshi T.E.,
- et al.