Using Glycosylated Hemoglobin to Define the Metabolic Syndrome in United States Adults Open Access

Data from the cross-sectional U.S. National Health and Nutrition Examination Survey (NHANES) 2003–2006 were used in initial analysis (6). For confirmation, we used the cross-sectional data from NHANES 1999–2002. All participants gave informed consent, and the study received approval from the Centers for Disease Control and Prevention Institutional Review Board.

MetS was defined using the consensus criteria in 2009 (4). Under this definition, a person has MetS if he or she meets three or more of the following criteria: 1) central obesity, defined using ethnic-specific cut points of waist circumference, 2) triglycerides ≥150 mg/dl (1.7 mmol/l), 3) HDL cholesterol <40 mg/dl (1.0 mmol/l) in men and <50 mg/dl (1.3 mmol/l) in women, 4) blood pressure ≥130/85 mmHg or on antihypertensive medication, or 5) raised blood glucose, defined as FPG ≥100 mg/dl (5.6 mmol/l) or on antidiabetic medication. For non-Hispanic whites and blacks, and people of other races and mixed races, the cut points for waist circumference were ≥94 cm in men and ≥80 cm in women. For Mexican Americans and other Hispanics, the cut points were ≥90 cm in men and ≥80 cm in women. In a separate analysis, MetS was defined using the NCEP criteria, which were the same as the consensus criteria in 2009 (4), except that central obesity was defined as waist circumference ≥102 cm in men and ≥88 cm in women (1,3). The uses of GHb ≥5.7% or FPG ≥100 mg/dl in the definition of the glycemic component of MetS were compared. Agreement between two definitions was defined as the percentage of participants who were classified the same under both definitions (7,8).

The laboratory methods have been described in detail elsewhere (6,8,,–11). Data on GHb and FPG were adjusted so that measurements across survey periods could be combined (6). History of cardiovascular diseases was obtained from self-reported questionnaires. Statistical analysis was performed using the complex samples function of SPSS version 15.0 (SPSS, Chicago, IL). Fasting sampling weights were used in all analyses to adjust for oversampling and nonresponse bias and to approximate the distribution to the U.S. population in the year 2000.

After excluding pregnant women and subjects with missing data in BMI, GHb, and the five components of MetS, there were 3,551 and 3,412 participants aged ≥20 years in NHANES 1999–2002 and 2003–2006, respectively, who had fasted for 8–24 h.

As shown in Table 1, in NHANES 2003–2006, the use of GHb alone resulted in a lower percentage of people meeting the glycemic criteria of MetS compared with the use of FPG alone, with 74.9% agreement between these two definitions. The use of GHb alone also resulted in a lower prevalence of MetS compared with the use of FPG alone (Table 1). The use of both GHb and FPG resulted in an insignificant increase in the prevalence compared with the use of FPG alone (41.1 vs. 38.8%, P = 0.200). There was 91.3% agreement between the use of GHb alone and the use of FPG alone. The agreement was good irrespective of age, sex, race/ethnicity, and BMI. The same trends were found in NHANES 1999–2002 or when the NCEP definition was used (see supplementary Tables A1 and A2, available in an online appendix at http://care.diabetesjournals.org/cgi/content/full/dc09-0190/DC1).

Because there is controversy whether the diagnosis of MetS conveys additional meaning in subjects with diabetes who should already be aggressively treated due to high cardiovascular risk, a subgroup without diabetes (non-DM) was also examined. Similar to the overall cohort, in the non-DM group, the use of GHb alone resulted in lower prevalence of MetS compared with FPG alone, with 90.6% agreement (Table 1). Importantly, in this non-DM subgroup, only the use of GHb alone, but not FPG, resulted in significant association with cardiovascular diseases (odds ratio 1.45, P = 0.005) when the consensus criteria in 2009 was used to define MetS (supplementary Table A3).

The controversy regarding the definition of MetS has been addressed recently in a joint scientific statement (4). GHb reflects the average blood glucose level over several months, and its measurement does not require a fasting blood sample. In this study, we demonstrated that there was good agreement between GHb and FPG in identifying individuals with MetS, despite only a moderate agreement (∼75%) between GHb and FPG in defining raised blood glucose. The components of MetS are inter-correlated; therefore, a certain degree of inaccuracy or fluctuation in one component is tolerated and does not result in misclassification. The agreement between GHb and FPG in the definition of MetS is good in different subgroups. We can therefore confidently conclude that using GHb instead of FPG to define MetS is feasible. This is true at least for Americans, based on the most up-to-date data on a nationally representative sample of Americans, and was confirmed using historical data. It remains to be seen if our conclusions are also applicable to Asians, among whom the prevalence of raised blood glucose is likely to be different.

The current cut point of GHb identifies a slightly smaller group of people as having MetS. However, it also identifies subjects at high risk for cardiovascular diseases, even in those without diabetes, when the consensus criteria in 2009 are used to define MetS. Whether GHb results in better risk stratification needs to be investigated in large prospective studies.

No potential conflicts of interest relevant to this article were reported.

K.L.O. researched data, contributed to discussion, and wrote the manuscript. A.W.K.T. and K.S.L.L. contributed to discussion and reviewed/edited the manuscript. S.S.C. and P.C.S. contributed to discussion. B.M.Y.C. contributed to discussion, wrote the manuscript, and reviewed/edited the manuscript.