Prevalence of the Metabolic Syndrome Among U.S. Workers

Data from the 1999–2004 National Health and Nutrition Examination Survey (NHANES), a multistage stratified complex design survey of a representative sample of the entire U.S. civilian population conducted by the National Center for Health Statistics (NCHS), was analyzed. In brief, trained interviewers and laboratory technicians conducted in-person interviews, performed physical examinations, and collected urine and blood samples either at mobile examination centers or at home (6). The response rates for participants interviewed in the NHANES surveys ranged from 79 to 84%, whereas the response rates for the participants examined ranged from 76 to 80% (6). Individuals who reported being employed and who had occupational group data, were ≥20 years, and were not pregnant were included in the analyses (n = 8,498).

The presence of the metabolic syndrome was based on the modified version of the definition recommended in 2001 by the Third Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (7,8). Metabolic syndrome was a dichotomous variable defined to be present or not based on having at least three of the following five criteria: 1) blood pressure ≥130/85 mmHg or receiving treatment for hypertension, 2) HDL cholesterol <50 mg/dl if a woman and <40 mg/dl if a man, 3) triglyceride level of ≥150 mg/dl, 4) waist circumference of >102 cm if a man or >88 cm if a woman, and 5) self-reported diabetes (9).

Employment status was based on the question “Did you work last week?” Occupational classifications were based on the 40 NCHS occupational codes (10) that appear in the NHANES data file. These variables were collapsed into 13 NCHS occupational groups. The collapsing of the 40 occupational groups into 13 occupational groups is the method used in all NCHS surveys, including the National Health Interview Survey with the occupational groups originally based on the more detailed U.S. Census Standard Occupation Classification System occupational groups (10,11). Table 2 shows where each of the 40 occupational groups falls within the 13 broader occupational groups.

Analyses were completed using SUDAAN (version 8.0) to take into account sample weights and design effects (12). The unadjusted and age-adjusted prevalence estimates for meeting the criteria for the metabolic syndrome were determined among workers aged ≥20 years. For unadjusted and age-adjusted prevalence estimates, all 40 occupational groups available in the NHANES data file were used. However, given the small sample size of workers in certain occupational groups, only 13 occupational groups were used for the logistic regression analyses. Occupation-specific prevalence estimates of metabolic syndrome were considered significantly “higher” than the “overall” sample prevalence rate if the occupation-specific prevalence was above the upper bound of the 95% CI for the overall sample. This is a variation on the method of testing a one-sample difference in proportions considering the overall sample as the population proportion (13).

Simple and multiple logistic regression analyses were then conducted with meeting criteria for the metabolic syndrome as the dependent variable (yes vs. no). Multiple logistic regression analyses adjusted for sex (male vs. female), age (in years), race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, and other), education (less than high school, high school education or equivalent, and greater than high school), health insurance (none vs. insured), BMI (underweight/normal, overweight, and obese), smoking status (nonsmoker, former smoker, and current smoker), alcohol drinking status (abstainer vs. drinker), and physical activity (none, moderate, and vigorous). An α level of 0.05 was used to determine statistical significance. This study was approved by the University of Miami Human Subjects Committee.

The prevalence of metabolic syndrome stratified by worker sample characteristics is shown in Table 1 (n = 8,457); the subgroup with the highest prevalence of metabolic syndrome was obese workers (42.5%), followed by workers aged ≥65 years (32.1%). Unadjusted and age-adjusted prevalence estimates for each of the 40 occupational groups are presented in Table 2.

The overall unadjusted prevalence estimate for all workers was 18.7% (95% CI [17.4–20.0%]), whereas the age-adjusted estimate was 20.6% [18.9–22.3%]. Occupations with the highest unadjusted prevalence for meeting criteria for the metabolic syndrome (all significantly higher than the prevalence for the overall sample) were “other transportation and material occupations” (33.1% [23.1–45.0%]), followed by “farm operators, managers, and supervisors” (27.4% [15.7–43.3%]), and “motor vehicle operators” (26.4% [21.2–32.2%]). The lowest unadjusted prevalence for meeting criteria for the metabolic syndrome was found among “writers, artists, entertainers, and athletes” (6.9% [3.6–12.8%]), followed by “waiters and waitresses” (7.6% [3.3–16.5%]) and “construction trades” workers (11.9% [8.4–16.6%]).

There was not much difference in the prevalence of meeting criteria for the metabolic syndrome after adjustment for age. However, the order or ranking of occupations with the highest prevalence did differ to some degree. For example, “other transportation and material occupations” and “motor vehicle operators,” the two occupations falling within the group of “transportation/material moving” were no longer the occupational groups with the highest prevalence for meeting criteria for the metabolic syndrome. After adjustment for age, occupations with the highest prevalence of the metabolic syndrome (all significantly higher than the prevalence for the overall sample) now included “miscellaneous food preparation and service occupations” (31.1% [95% CI 19.6–45.4%]), followed by “farm operators, managers, and supervisors” (29.8% [13.8–52.9%]), and “health service occupations” (26.6% [19.4–35.3%]). The lowest age-adjusted prevalence of the metabolic syndrome was documented in “writers, artists, entertainers, and athletes” (8.5% [4.5–15.4%]), “engineers, architects, scientists” (9.2% [6.2–13.6%]), and “health diagnosing, assessing, and treating” workers (11.8% [7.2–18.7%]).

The logistic regression analyses adjusting for demographics and potential confounders showed that “transportation/material moving” workers relative to “executive, administrative, managerial” professionals were significantly more likely to meet the criteria for the metabolic syndrome (odds ratio 1.70 [95% CI 1.15–2.52]) (Table 3). Among all U.S. workers, other participant characteristics with significantly greater odds of meeting criteria for the metabolic syndrome included older age (1.03 [1.03–1.04]) and being overweight (5.63 [3.80–8.35]) or obese (25.94 [18.08–37.23]) relative to underweight or normal weight. Lower odds for metabolic syndrome included being non-Hispanic black (0.48 [0.36–0.65]) relative to non-Hispanic white, alcohol consumer relative to non–alcohol consumer (0.78 [0.64–0.97]), being a former smoker relative to a never smoker (0.81 [0.67–0.97], and doing vigorous physical activity relative to no physical activity (0.63 [0.53–0.75]).

This is the first nationally representative study of U.S. workers to estimate the prevalence of metabolic syndrome in various occupational groups. In both unadjusted and age-adjusted analyses, we found a threefold difference in the prevalence of metabolic syndrome across occupational groups, with the greatest unadjusted prevalence among “other transportation and material occupations” and age-adjusted prevalence among “food preparation and food service workers.”

Differences in the prevalence of metabolic syndrome by occupation are likely to be strongly influenced by differences in the prevalence of obesity (14). Interestingly, even after adjustment for potential confounders including obesity, older age, sex, race/ethnicity, education, physical activity, alcohol consumption, and smoking, “transportation and material moving workers” showed statistically significant greater odds for meeting the criteria for metabolic syndrome compared with other workers. This finding is consistent with several studies that have found transportation workers (such as truck drivers) to have a higher prevalence and incidence of cardiovascular disease, including heart disease and stroke (15,16). A potential explanation for the relationship between transportation work and meeting the criteria for the metabolic syndrome could be more irregular work schedules and shift work, sleep problems, and job stress, as these factors have been associated with metabolic syndrome (3,–5,17,18); of note, each of these occupational factors is more prevalent among transportation workers relative to other occupational groups (16,19,20). Additional research is needed to understand the relative role that these occupational risk factors play in influencing metabolic syndrome prevalence rates across occupational groups, as well as occupation exposures, which may be unique among “transportation/material moving” workers.

The present study had several limitations, such as its cross-sectional design, which did not allow for causal inferences. Another limitation was the lack of fasting glucose values for determination of metabolic syndrome status among all NHANES study participants, which could have led to an underestimate of the prevalence of metabolic syndrome in this study. However, sensitivity analyses were performed in the subsample (one-third of the total NHANES sample) that did have the fasting blood glucose data needed for defining metabolic syndrome (i.e., with having a metabolic risk factor of having self-reported diabetes or a fasting blood glucose measurement of ≥100 mg/dl). Although not statistically significant, the results were similar in terms of direction of the estimates with use of the previous definition (i.e., self-report of diabetes only). Details about working conditions or work characteristics were not available in NHANES. Thus, we were unable to examine correlates of work schedule, sleep patterns and problems, and occupational stress on metabolic syndrome prevalence rates. Furthermore, data on type of occupation was only available in the continuous NHANES from 1999 to 2004, thereby limiting the sample size that would have been beneficial in looking at more specific occupational groups (i.e., 40 categories). Finally, given differences in survey design, it is not appropriate to merge NHANES III (1988–1994) data with data from the continuous NHANES (i.e., 1999 and forward).

In conclusion, our findings have implications for policy makers and employers. Given that studies have shown greater reports of missed work (21,22) and presenteeism (23) among U.S. individuals with the metabolic syndrome compared with individuals without metabolic syndrome independent of obesity, it would seem beneficial for occupational health advocates and employers to be aware of the prevalence of metabolic syndrome among their employees and the associated consequences. To offset such work implications, employers and occupational health advocates should introduce metabolic syndrome awareness, management, and preventive programs at the workplace, particularly in occupational groups in which the overall prevalence of metabolic syndrome is high. Thus, according to our findings, metabolic syndrome-related interventions appear to be most needed for “transportation and material moving” workers as well as for “farm operators, managers, and supervisors” and “miscellaneous food preparation and service occupations.” Given the greater odds of metabolic syndrome among “transportation/material moving” workers even after adjustment for potential confounders, future occupational health research should examine factors that may explain the higher likelihood of metabolic syndrome in this occupational group. Finally, the high prevalence of the metabolic syndrome among older workers (24), combined with the growing numbers of older adults in the U.S. workforce (25), may lead to an increasing number of workers with metabolic syndrome and co-occurring cardiovascular consequences unless effective prevention programs, particularly those implemented in worksites for higher prevalence occupations, are rapidly developed and implemented.

This research was supported in part by the National Institute on Occupational Safety and Health (grant R01-0H-03915).

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

E.P.D. researched data and wrote the manuscript. H.F. contributed to discussion and reviewed/edited the manuscript. L.E.F. contributed to discussion and reviewed/edited the manuscript. D.J.L. contributed to discussion and reviewed/edited the manuscript. E.G. contributed to discussion and reviewed/edited the manuscript. W.G.L. analyzed data and provided statistical advice. A.J.C.-M. contributed to discussion and reviewed/edited the manuscript. K.L.A. and K.E.M. contributed to discussion, provided statistical advice, and reviewed/edited the manuscript. S.L.C. contributed to discussion and reviewed/edited the manuscript. J.C.C. contributed to discussion and reviewed/edited the manuscript. T.C. contributed to discussion and reviewed/edited the manuscript.