The Influence of Metabolic Syndrome in Predicting Mortality Risk Among US Adults

Importance of Metabolic Syndrome Even in Adults With Normal Weight

Ting Huai Shi, BA; Binhuan Wang, PhD; Sundar Natarajan, MD, MSc


Prev Chronic Dis. 2020;17(5):E36 

In This Article


When we evaluated mortality risk by obesity and MetS categories, only the normal-weight and obese groups with MetS had a significantly higher hazard ratio than the reference group, adults with normal-weight and no MetS. Although the normal-weight group had lowest prevalence of MetS, it also had the highest hazard ratio. Analysis of cardiovascular mortality showed a significantly higher hazard ratio only for the normal-weight–MetS group, which suggests a strong effect of MetS on cardiovascular death in normal-weight adults. Cancer mortality was significantly higher in the overweight–MetS and obese–MetS groups, compared with the normal-weight–no-MetS group, which is consistent with a previous study that found strong associations between adiposity and risk for many types of cancers.[21]

The most likely explanation for the higher mortality in the normal-weight–MetS group is the influence of MetS through obesity-independent risk pathways. Although obesity is a known, common risk factor for MetS, MetS in the normal-weight population is likely due to factors independent of obesity. If these obesity-independent factors result in a more severe form of MetS, the normal-weight–MetS group would show a higher mortality rate than the obese–MetS group, whose mortality rate is attenuated by the less severe MetS caused by obesity. This phenomenon, sometimes known as a "collider stratification bias," is cited as a possible cause of the obesity paradox, a phenomenon that similarly describes a lower mortality risk in obese patients with diabetes or cardiovascular disease.[22]

Previous studies on normal-weight obesity attribute the cause of MetS in normal-weight adults to excess body fat percentage despite a normal overall weight. One study found that among participants with normal-weight BMI, those with a higher body fat percentage had a higher prevalence of abnormality in every MetS component.[23] Normal-weight "obese" people have a BMI of less than 25 kg/m2, but they have symptoms of metabolic obesity, such as low insulin sensitivity, high hepatic fat, and high triglycerides. The location of adipose tissue can also affect metabolic health. Increased visceral adipose tissue accumulation is more strongly associated with risk of metabolic disorders than subcutaneous adipose tissue because of its location in the abdominal cavity and its large role in endocrine and inflammatory secretion.[24] Therefore, the difference in adiposity of 2 people of similar weight can result in different susceptibility to insulin resistance and MetS, and this difference cannot be determined by BMI values.

Our findings do not mean that being obese with MetS is beneficial compared with being normal weight with MetS. Studies on the prognosis of overweight and obese patients show that risk of cardiovascular disease[25] inversely correlates with increasing weight.[26] Other studies on the effect of weight loss show a significant improvement in all risk factors and symptoms of MetS after weight loss.[27]

The results of 2 previous studies that examined mortality rates in MetS and BMI categories using NHANES III (1988–1994) data found results that are contradictory to the results of our study. Kuk and Ardern found higher HRs for the obese–MetS group, the obese–no-MS group, and the overweight–MetS group, but not for the normal-weight–MetS group when they used the normal-weight–no-MetS group as the reference.[12] Durward et al found that only obese adults with MetS had a higher hazard ratio compared with the normal-weight–no-MetS group.[13] However, although neither study found a significantly higher mortality risk in the normal-weight–MetS group compared with the normal-weight–no-MetS group, the HRs were 1.25 and 1.4, respectively. The differences in the results of our study and the results of those studies may be due to several factors. One, the NHANES III sample size was less than half of the Continuous NHANES sample size used in our study. Kuk and Ardern included 6,011 participants, and the Durward et al study included 4,373 participants, of whom only 77 were in the normal-weight–MetS group. The lack of significant differences in HRs between the normal-weight–MetS group and the normal-weight–no-MetS group in both studies was probably due to their small sample sizes. Two, the age-adjusted prevalence of MetS increased significantly from 29.2% in NHANES III (1988–1994) to 34.2% in Continuous NHANES (1999–2006).[5] Three, Kuk and Ardern included only 8 years of mortality follow-up data, resulting in a lower mortality rate than that observed in our 15 years of follow-up. Four, we used a more stringent definition of MetS than Kuk and Ardern used. Their definition excluded the criterion of central obesity and required only 2 conditions for diagnosis, which may have diluted the potent effect of MetS on all-cause mortality. Similar to our findings, Sahakyan et al found higher mortality in the normal-weight-with-central-obesity group than in their normal-weight-no-central-obesity group.[11] Other studies that examined groups of people of primarily European descent stratified by BMI and metabolic health criteria — usually MetS or insulin resistance — found similarly higher risks for cardiovascular disease and diabetes in the normal-weight but metabolically unhealthy group,[6–11] whereas 2 other studies found contradictory results.[28,29]

The main strength of our study is the use of data from the Continuous NHANES, which rigorously measures all components of MetS, has data on a large number of people, and oversamples racial/ethnic minority groups, the elderly, and the poor. These weights, strata, and cluster variables were included in all analyses and allow our results to be generalizable to the US population. Therefore, in addition to usually studied population subgroups, our study included information on a large proportion of African Americans and Hispanic Americans. These populations are mostly absent in European studies, where normal-weight MetS was found to be a significant risk factor for cardiovascular disease and diabetes. NHANES also provides the statistical power necessary to answer the important question on the influence of MetS on normal-weight people. Furthermore, the Continuous NHANES is more recent and up to date than NHANES III, the database used in similar previous studies, so the data are more directly applicable to current clinical and public health practice.

Our study has several limitations. One limitation is the use of BMI to define weight groups. Many researchers argue that BMI is not a good indicator of weight status, because it does not differentiate among body compositions.[26,30] Another limitation is that we lacked data from longitudinal time points. We have no information on how the characteristics of each participant, such as BMI, MetS, and other risk factors for mortality, changed from the time of the NHANES examination. The weight and metabolic condition of participants over multiple points as time-varying covariates would be needed to achieve a more comprehensive analysis than the one provided here.

Our study found that the normal-weight–MetS group had the highest risk of mortality among MetS and obesity categories — a risk that has not been previously identified in US adults. As such, greater attention must be given to normal-weight people who have MetS to provide early treatment and prevent future complications. In addition, the obese–MetS group had high mortality risk, particularly cancer mortality, when compared with the normal-weight–no-MetS group. Our study re-emphasizes that weight loss should continue to be encouraged among people who are obese.