Self-reported type 2 diabetes Mellitus is associated with abdominal obesity and poor perception of health in shift workers

Objective To investigate factors that are associated with type 2 diabetes Mellitus in shift workers of a slaughterhouse in Southern Brazil. Methods This cross-sectional study included 1,194 18to 50-year-old workers of both sexes. The presence of type 2 diabetes Mellitus was self-reported and confirmed by the use of hypoglycemic drugs or insulin. The independent variables were sex, age, skin color, marital status, education level, family income, leisure time physical activity, smoking, and self-reported health and nutritional status (body mass index and waist circumference). Multivariate analysis was performed from an a priori conceptual model. 776 | C FRÖHLICH et al. Rev. Nutr., Campinas, 29(6):775-783, nov./dez., 2016 Revista de Nutrição http://dx.doi.org/10.1590/1678-98652016000600003 Results The prevalence of diabetes was 1.3% (95%CI=0.6-1.9). Type 2 diabetes Mellitus was associated with poor or regular self-reported health (OR)=3.72; 95%CI=1.28-10.78) and level II abdominal obesity ≥102 for men and ≥88 for women (OR=5.76; 95%CI=1.07-29.10). Conclusion The prevalence of type 2 diabetes Mellitus was low. Moreover, the study evidenced the importance of using waist circumference to surveil and screen for metabolic diseases, such as type 2 diabetes Mellitus, and to monitor the low quality of life in the study individuals given the poor self-perceived health of workers with the said disease.


I N T R O D U C T I O N
Type 2 Diabetes Mellitus (DM2) is a metabolic disease characterized by the lack or ineffectiveness of insulin.Studies show that diabetes increases the risk of developing vascular diseases, such as myocardial infarction and stroke 1-4 .
According to the International Diabetes Federation (IDF), there are 382 million diabetics in the world, exceeding, in 2012, the projections for 2030.Brazil ranks fourth country in the occurrence of diabetes, with a prevalence of 6.9% 5 .However, there are no nationwide estimates of DM2 prevalence in Brazil, only selfreported measures have been used in nationwide studies.
The etiology of DM2 is multifactorial.The causes associated with its occurrence include sociodemographic factors, such as age, skin color, education level, and income; behavioral factors, such as eating habits, tobacco use, and physical inactivity; and body adiposity [6][7][8] .In addition, the http://dx.doi.org/10.1590/1678-98652016000600003workplace may function as an indirect contributing factor due to its characteristics, such as long working hours, stressing, insufficient occupational physical activity, and shift work 9,10 .The number of studies that investigated the occurrence of metabolic disorders in workers have increased [11][12][13][14] .
Type 2 diabetes Mellitus symptoms can interfere on workers' quality of life and productivity, and this is an important occupational health aspect.This study investigated the factors associated with DM2 in shift workers of a poultry processing plant in southern of Brazil.

M E T H O D S
A cross-sectional study was conducted with a representative sample of 1,194 shift workers aged 18-50 years.The subjects worked in the production line of a round-the-clock poultry processing plant in southern Brazil.
The sample size was calculated to assess the association between work shift and obesity 12 based on the following presuppositions: confidence level of 95%; statistical power of 80%; unexposed: exposed ratio of 1:2 and risk ratio of 1:30.Sample size was increased by 10% to compensate for possible losses or refusals, so 1,125 workers were needed.All workers (1,270) who resided in the municipality where the plant is located and two nearby municipalities, working in one of the three production sectors (cutting room, thermoprocessed foods, and evisceration) were considered eligible for the study.Pregnant workers and those on leave for more than 10 days during the study period were excluded.There were 76 losses and refusals, resulting in 1,194 workers who remained in the study.
Standardized, coded, and pretested questionnaires collected demographic, socioeconomic, behavioral, and outcome data.The presence of DM2 was assessed by the question: "Has any doctor ever told you that you have diabetes Mellitus?"An affirmative answer to this question was confirmed by the use of hypoglycemic drugs or insulin.
The study demographic and socioeconomic variables were sex (men/women), age (collected as a continuous variable sorted into three categories: 18-29 years, 30-39 years, and ≥40 years), skin color (self-reported by the interviewees and categorized as white or other), education level (collected as a continuous variable and categorized as ≤4 years, 5-10 years, and ≥11 years), and family income (total income of each family member in the last month and categorized as tertiles (low, medium, high).The study behavioral variables were leisure-time physical activity (physically active: ≥150 min/week and inactive: <150min/week) 15 , smoking ("never smoked", "ex-smoker", and "smoker"), and selfrated health status (poor, regular, good, very good, and excellent), later categorized as good self-rated health (excellent, very good, and good) and regular/poor self-rated health (regular and poor).
The workers' nutritional status and abdominal obesity were investigated by anthropometry.Nutritional status was defined by body mass index (BMI=kg/m²) and categorized as normal weight (≤25.0 kg/m²), overweight (25.0-29.9kg/m 2 ), and obese (≥30.0 kg/m 2 ).Weight was measured by the Fantasy Sunrise scale, digital model 2096PP (Toledo, São Paulo, Brazil) with capacity of 130 kg and accuracy of 100 g.Height was measured by a stadiometer (SECA Bodymater 208, Hamburg, Germany) with a measuring range of 0-200 cm and accurate to 01 mm.

Abdominal obesity was assessed measuring
Waist Circumference (WC) at the midpoint between the lowest rib and iliac crest.The cut-off points were: appropriate (<94 cm and <80 cm for men and women, respectively), Level I (WC ≥94 cm to <102 cm for men and 80≥ to <88 cm for women), and Level II (≥102 cm and ≥88 cm for men and women, respectively) 16  .The crude and adjusted odds ratios and respective 95% confidence intervals (95%CI) were estimated by logistic regression.The prevalence of DM2 was estimated to be lower than 6%, allowing the use of logistic regression without overrating effect measures.For the adjusted analysis, a conceptual model of analysis was used in which the variables were grouped from the more distally to the more proximally related to type 2 diabetes Mellitus.Therefore, the first level included demographic and socioeconomic variables, the second level included behavioral variables, and the third level included anthropometric variables.Only variables with p<0.20 in the crude analysis were included in the adjusted analysis.Afterwards, the variables included in the adjusted analysis were controlled for all the other variables of the same level, and those that presented p<0.20 were maintained in the next level of analysis.
Table 2 shows the crude and adjusted odds ratios.After the adjustments, the variables that remained statistically associated were physical activity, self-rated health, and abdominal obesity.Workers with level II waist circumference (≥102 for men and ≥88 for women) were almost six times more likely to have DM2 than those with normal waist circumference (<94 cm for men and <80cm for women).Active workers and workers with regular/poor self-rated health were 3.7 times more likely to have DM2 than inactive workers or workers with good self-rated health.

D I S C U S S I O N
The main finding of our study is the low prevalence of DM2 among workers, although it is strongly associated with abdominal obesity and poor or regular self-rated health, regardless of nutritional status.
The 1.3% prevalence of DM2 found among workers was lower than that found by a Brazilian nationally representative study (6.9%) 5 , and similar to others found among workers of 1.2% 17 and 1.7% 18 .This result may be explained by the healthy worker bias, i.e., relatively healthy people are or remain employed, whereas those with health problems are unemployed, retired, and/or unable to work.It should be pointed out that the symptoms of DM2 (polyuria, blurred vision, and others) may undermine productivity in the workplace.Therefore, the occurrence of DM2 may be a limitation for certain types of work that demand physical strength and long working hours, such as work performed in a poultry processing plant.Workers in large companies are periodically examined, which favors the diagnosis and prevention of chronic Noncommunicable Diseases (NCD).Moreover, these companies also allow workers to have at least one meal in the factory's canteen, thus providing a balanced energy and nutrient supply in at least one meal a day.Finally, self-report measures may underestimate DM2 prevalence 19 .All these aspects may contribute to a lower prevalence of DM2 in the workers of our sample.
The practice of physical activity as a protective factor against DM2 is well documented http://dx.doi.org/10.1590/1678-98652016000600003 in the literature [20][21][22] .A systematic review evaluated the evidence from ten prospective studies, including a total of 301,221 adults and 9,367 incident cases, and follow-ups ranging from four to 17 years.The review found that moderate physical activity protected against the risk of type 2 diabetes 20 .However, in our study, DM2 was associated with the practice of 150 minutes or more of physical activity per week.Our finding may probably be explained by reverse causality.Physically-active workers may have been physically inactive for almost all their lives but changed their Note: * Fisher's Exact Test for heterogeneity of proportions (categorical variables) and linear trend (ordinal variables); ** Normal: ♀ <80cm and ♂ <94cm, Level I: ♀ ≤80 cm to <88 cm and ♂ ≤ 94 cm to <102 cm, Level II: ♀ ≥88 cm and ♂ ≥102 cm.95%CI: 95% Confidence Interval; BMI: Body Mass Index.
The workers who rated their health as regular/poor were more prone to having DM2 than those who rated their health as good.This result is expected, since the presence of DM2 may cause many symptoms, such as thirst, polyuria, blurred vision, weight loss, and hyperphagia.In its most severe forms, it may cause ketoacidosis or nonketotic hyperosmolar syndrome, drastically diminishing the quality of life of the affected individuals 23,24 .These associations have already been observed in other studies 25,26 .
Abdominal obesity was associated with DM2 regardless of workers' BMI.Metabolic complications associated with obesity were not only related to excess weight, but also, and mainly, to the distribution of body fat.In this sense, studies have shown that localized fat in the central region of the body, particularly visceral adiposity, has a greater influence on the development of NCD, such as DM2, as it is a reflection of a pathological state associated with metabolic diseases [27][28][29][30] .Waist circumference is the anthropometric measure most strongly correlated with the quantity of visceral adipose tissue [28][29][30] .Thus, occupational health professionals may incorporate waist circumference measurement in their routine evaluations as it is an important marker of cardiometabolic risk.Some limitations of this study, such as reverse causality and the healthy worker bias, have already been discussed.Another possible limitation is the use of self-reported measures for measuring the outcome.However, self-reported measures of NCD have been widely used in prevalence studies.In Brazil 1 important nationally representative epidemiological studies, such as the Vigilância de Fatores de Risco e Proteção para Doenças Crônicas por Inquérito Telefônico (VIGITEL, Telephone Survey on Chronic Noncommunicable Disease Protective and Risk Factors, Brasília, Distrito Federal, Brazil) and the Programa Nacional por Amostra de Domicílio (PNAD, National Household Sample Survey), have employed the same methodology 5,31 .It should also be highlighted that, based on the sample size (n=1,194), a 1.5% prevalence of DM2 could be observed, considering a confidence level of 95.0%, an acceptable error of 1 percentage point, and a statistical power of 80.0% to detect a prevalence ratio of 3.8 between DM2 and its associated factors.Finally, a conceptual model was used in adjusted analysis; however, the variables included in this model explained only 16.0% of the outcome (pseudo-R2=0.16).

C O N C L U S I O N
The results of this study point to a greater prevalence of DM2 in workers with abdominal obesity, reinforcing it as an important predictive factor for this disease.In this sense, the efficiency and practicality of waist circumference as a screening measure for workers should be highlighted.In addition, most diabetic workers rated their health status as regular/poor, showing that the symptoms of DM2 decreases the quality of life of its carriers, which may result in loss of productivity in the work place and, therefore, exclusion from the labor market.Thus, in addition to preventive DM2 actions, workers who already have diabetes should receive special care in order to avoid the worsening of the disease, preserve their quality of life, and prevent exclusion from the labor market. /dx.doi.org/10.1590/1678-98652016000600003 A C K N O W L E D G M E N T S This study was sponsored by Conselho Nacional de Desenvolvimento Científico e Tecnológico.C FRÖHLICH receives a student grant from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior.MTA OLINTO and MP PATTUSSI received research productivity grants from Conselho Nacional de Desenvolvimento Científico e Tecnológico.C O N T R I B U T O R S All authors participated in all phases of the research.R E F E R E N C E S 1. Sarwar N, Gao P, Seshasai SR, Gobin R, Kaptoge S, Di Angelantonio E, et al.Diabetes mellitus, fasting ://dx.doi.org/10.1590/1678-98652016000600003Denmark).The statistical analysis was performed by the software Stata version 11.0 (Stata Corporation, College Station, Texas, United States) . Data were doubly entered into the software Epi Data version 3.1 (Epi Data, Odense, Rev. Nutr., Campinas, 29(6):775-783, nov./dez., 2016 Revista de Nutrição http

Table 1 .
Sample characteristics and prevalence of type 2 Diabetes Mellitus (DM2) according to sociodemographic, behavioral, and anthropometric characteristics of shift workers of a poultry processing plant in southern Brazil.(n=1,194).

Table 2 .
Crude and adjusted Odds Ratios (OR) and their respective Confidence Intervals (95%CI) for type 2 Diabetes Mellitus (DM2) according to sociodemographic, behavioral, and anthropometric characteristics of shift workers of a poultry processing plant in southern Brazil.(n=1,194).