Metabolic syndrome , insulin resistance and other cardiovascular risk factors in university students

1 Programa de PósGraduação em Saúde Coletiva, Universidade Federal do Maranhão (UFMA). Av. Barão de Itapary 227, Centro. 65020-070 São Luís MA. josebbarbosa@uol.com.br 2 Programa de PósGraduação em Ciências da Saúde, UFMA. São Luís MA Brasil. 3 Programa de PósGraduação da Rede Nordeste de Biotecnologia. Recife PE Brasil. Metabolic syndrome, insulin resistance and other cardiovascular risk factors in university students


introduction
Cardiovascular diseases are a serious public health problem in Brazil, currently being considered the main cause of death and disability worldwide 1,2 .Mortality rates from cardiovascular diseases (CVDs) have greatly increased in middle-income countries, such as Brazil 2 .According to the Ministry of Health, 30.7% of the total deaths in Brazil in 2011were caused by cardiovascular diseases 3 .
Although each cardiovascular risk factor in isolation has an impact on health, very often these risk factors are aggregated in individuals 4,5 .This aggregation of cardiovascular risk factors, such as abdominal obesity and insulin resistance, comprises the metabolic syndrome (MS) 2 .MS increases mortality from CVD by 2.5 and is recognized worldwide as a major public health problem 6 .
This set of cardiovascular risk factors in the adult and elderly population is a common finding well described in the literature.However, studies have shown the increasingly occurrence of cardiovascular events and risk factors among adolescents and young adults 1,7,8 .Few data regarding the prevalence of MS, insulin resistance (IR) and other risk factors for CVD in young adults are available in Brazil.Most existing studies have usually been performed in small samples from selected institutions 1,9,10 .In the few available population-based studies the prevalence of MS varied from 15.8% in Vitória 11 to 24.4% in the Federal District 12 among people aged 25-34 years.In young adults aged 23-25 years from Ribeirão Preto, the prevalence of MS was 7.6%, whereas the prevalence of insulin resistance was 13.9% 13 .
Due to the scarcity of studies that addressed MS, IR and other cardiovascular risk factors in young adults in Brazil and especially in the Northeast, the region in which these events are on the increase 12 , this study was carried out with the objective of estimating the prevalence of MS, IR and other modifiable and non-modifiable risk factors for cardiovascular diseases among college students in a population-based sample in São Luís, Maranhão, Brazil.Differences in cardiovascular risk factors between men and women and between public and private institutions were also assessed.

Methods
The present cross-sectional population-based study was conducted in three public and six private Higher Education Institutions (HEIs) in the city of São Luís, Maranhão, Brazil.The data are from a probabilistic sample of college students, and were collected from August 2011 to October 2012.

Sampling
Nine HEIs were included in the sampling frame.Together, they accounted for 95% of all university students in the city.Institutions with small numbers of students were not included.Sampling was carried out in clusters in two stages.In the first stage disciplines were selected and in the second stage students were selected.The sample was stratified into two groups: public and private institutions.In each institution a list of all disciplines offered was obtained.From this list in each stratum a simple random sampling of disciplines was carried out, with probability proportional to the number of students in each institution in relation to the total number of students in all universities.Then, in each selected discipline, 12 students were randomly selected to participate in the study.The probability of selection for each student was conditional on the number of disciplines that he/she was attending.The higher the number of disciplines the student was attending the higher the probability of he/she being selected.Thus, probabilities of selection were corrected for multiplicity, considering the number of disciplines taken by each student.
The estimated sample size was 1276 students.This sample size allowed estimating a prevalence of approximately 50% with a margin of error of 3% and 95% confidence.With this sample size it was possible to detect an 8% difference in the prevalence of metabolic syndrome (estimated at 10%) between the exposed and unexposed individuals, assuming a 5% probability of type I error, settling at 80% the study power and considering a design effect of 2. The final sample was of 968 students.Losses amounted to 24.1% due to refusals or absence of the student in the classroom on the day of interview.

Data collection
Data were obtained through interviews with the students by means of a standardized questionnaire.The following data were collected: age, sex, education of the household head (College/ University degree, high school, primary or middle school), Brazilian economic classification (based on the possession of goods and education of the family head, classified into A to E, with A being the most educated and wealthy and E the least), university type (public or private), marital status, family income, self-reported skin color (classified according to the Brazilian Institute of Geography and Statistics -IBGE 11 -into black, brown/mullato and white), religion, possession of health insurance, participation in the ProUni (University for Everyone Program), tobacco and alcohol use in the last month.Smoking was considered the consumption of at least one cigarette in the last month.Fat intake was assessed by the Block score, and considered high when > 27 14 .
Blood pressure and anthropometric measurements were also carried out and a blood sample was drawn for laboratory tests.Of the 968 participants, 590 (61.0%) attended blood collection.
Blood pressure was measured twice at five-minute intervals and only the lowest value was considered in data analysis.We used digital Omron® automated devices with different cuff sizes, following the VI Brazilian Guidelines on hypertension 15 .
To check participant's weight and body fat percentage (by bipolar bioelectrical impedance) the Tanita® (BC533) portable scale was used.The procedure was performed with the student standing barefoot on the metal surface conductive equipment according to the manufacturer's guidelines.Percent body fat cutoff points followed guidelines by Lohman 16 .
The measurement of height was held in the Alturaexata® brand stadiometer with the subject standing barefoot, heels together, back straight and arms extended alongside the body.Body Mass Index (BMI) was calculated according to the formula: BMI = weight (kg)/height (m)².BMI classification was performed according to the cutoff points proposed by the World Health Organization in 1998 17 .
Waist circumference (WC) was measured with the student standing, using a non-stretchable tape at the midpoint between the iliac crest and the outer face of the last rib, and readings were made at the time of expiration 18 .WC values greater than 102 cm for men and 88 cm for women (National Cholesterol Education Program -Adult Treatment Panel III -NCEP ATPIII) and 90 cm for men and 80 cm for women (International Diabetes Federation-IDF and Joint Interim Statement-JIS) were considered indicative of risk for cardiovascular diseases 19 .
Hip circumference (HC) was obtained in the region of the largest circumference between the waist and the thigh, with a non-stretchable tape.With data from WC and HC, the waist-hip ratio (WHR) was calculated.WHR values greater than 0.90 for men and 0.85 for women were considered of risk 18 .
The waist-to-height (WHtR) was calculated based on measurements of WC and height.WHtR values greater than or equal to 0.52 for men and 0.53 for women were considered of risk for CVD 18 .
Fasting glucose and lipids (Triglycerides, HDL-c) were measured in the ADVIA 1650 equipment (Bayer CO, USA) and fasting insulin in the IMMULITE 2000 (Siemens) unit.
For the calculation of IR blood glucose levels and fasting insulin were used.IR was measured by the Homeostasis Model Assessment -Insulin Resistance (HOMA-IR) according to the following formula: HOMA-IR = insulin (U/mL) x (glucose mg / dL ÷ 18) ÷ 22.5.HOMA-IR estimates insulin sensitivity and beta cell function from the pancreas 20,21 .Insulin resistance was defined as HOMA-IR > 2,7 22 .
The diagnosis of MS was completed by the NCEP ATPIII, when attended at least three of the five following criteria: triglycerides > 150 mg / dl; HDL-C < 50 mg/dl in women and < 40 mg/dl in men, fasting glucose > 100 mg/dl, waist circumference > 102 cm for men and > 88 cm for women and blood pressure > 130x85 mmHg.MS according to IDF was defined when waist circumference > 90 cm for men and > 80 cm for women and two of the other above mentioned components were present.MS was defined according to JIS as the presence of three of the above five mentioned criteria with a lower cutoff point for the WC: > 90 cm for men and > 80 cm for women 19 .Following IDF recommendations adult criteria was used for adolescents ≥ 16 years of age 23 .
Physical activity was assessed by applying the International Physical Activity Questionnaire (IPAQ short version).Individuals were classified into active or sedentary according to proposed IPAQ guidelines 24 .

Statistical analysis
For descriptive analysis, the mean and standard deviation were calculated for quantitative variables with normal distribution and the median and 25th and 75th percentiles for those with non-normal distribution.Percentages were presented for qualitative variables.Data were processed and analyzed in Excel and Stata 12.0.Differences in prevalence by sex were tested by the chi-square and differences in quantitative variables were verified using the Student t test or the nonparametric Mann-Whitney test, when appropriate.The significance level was set at 0.05.
The probability of selection for each student was calculated for each sampling stage.For the first stage it was the ratio between the number of students in each class divided by the total number of subjects offered in each university.For the second stage it was the number of students interviewed, divided by the number of students attending each discipline.These two probabilities were then multiplied by each other and then by the number of disciplines taken by each student.
In the statistical analysis we took into account stratification according to university type (public or private) and the design effect through svy commands in Stata.Data were weighted by the inverse probability of selection of each participant.Since 39% of those who participated did not attend blood collection, differences in percentage of losses according to some variables were evaluated using the chi-square test.To reduce selection bias the inverse probability weighting procedure also took non-response into account.Variables that were associated with participation were identified in two-way tables by chi-square tests.Then, probabilities of participation in laboratorial exams were estimated by logistic regression including family income, schooling of the head of family, stratum (public or private university), religion and health insurance as predictors.The final weight w i for individual i used in the analysis of laboratorial exams was then the product of two weights: the inverse of the probability of being sampled and the inverse of the probability of participation in blood collection.

ethical aspects
This study was approved by the Research Ethics Committee of the University Hospital of the Federal University of Maranhão.Each participant signed the Informed Consent Statement Form.

results
The sample consisted of 968 students, 614 females (62%) and 354 males (38%).Age varied from 16 to 62 years.The median age was 22 years for men and 23 years for women.
A higher percentage of students had family incomes from 2 to 4 minimum wages (35.5%), had household heads with secondary education (56.3%), belonged to the classes A/B (45.8%) or attended private universities (65.6%).Fifty-two percent reported being brown/mullato, 82.1% were living without a partner, 59.2% profess the Catholic religion, and 53% had no health insurance.A small percentage (8.9%)participated in the ProUni (University for Everyone Program) (Table 1).
Measures of central tendency (mean or median) of weight, height, body mass index, waist circumference, hip circumference, waist-to-hip ratio, waist-to-height ratio, systolic and diastolic BP, fasting glucose and triglycerides were higher for men compared to women (P < 0.01).Fasting insulin, body fat percentage and HDL-cholesterol were higher among women than men.Age (p = 0.657) and HOMA-IR (p = 0.196), presented no statistically significant difference when comparing men and women (Table 2).
According to BMI, 29.5% were overweight and 8.7% obese.The prevalence of obesity was higher for men (12%) than for women (6.7%).Men also had a higher prevalence of overweight (39.1%) than women (23.7%) (p < 0.001).High waist-to-hip ratio was observed in 21.1% and the waist-to-height ratio was increased in 27.9%, with higher values for men.By body fat percentage 23% were obese, with no difference between sexes (p = 0.525) (Table 3).
Higher prevalences of physical inactivity (p = 0.019), obesity according to BMI (p = 0.024), high waist-to-hip ratio (p = 0.001), increased waist-to-height ratio (p <0.001), increased body fat percentage (p = 0.001) were observed among students from private compared with those attending public institutions (Table 5).Increased waist circumference by the NCEP and the IDF criteria (p = 0.007 and p < 0.001, respectively) and high triglycerides (p = 0.015) were also higher for students attending private institutions (Table 6).Prevalence of MS according to the NCEP and IR did not differ between students from public and private universities.However, according to the IDF and JIS criteria, students from private institutions had higher prevalences of MS compared to those from public ones (Table 6).
Table 7 compares characteristics of students who did or did not show up for collection of blood samples.Greater percentage of losses was observed among university students with higher family income (p = 0.034), greater education of the household head (p = 0.003), from private institutions (p < 0.001), who had health insurance (p = 0.012), from the catholic religion or from those who had no religion (p = 0.009).

Discussion
The main findings of this study were high prevalences of decreased HDL-C (61.2%), hypertension (27.1%) and metabolic syndrome (20.5% according to JIS).In addition, higher prevalence of cardiovascular risk factors were observed in male college students and in those attending private institutions.It was noteworthy that, in this predominantly middle and high-income young population with a median age of 22 years, the prevalence of insulin resistance was 7.3%.
In a study by Bustos and colleagues, performed in a young population with roughly the same age in Limache (L), Chile, and Ribeirão Preto (RP), Brazil 13 , high prevalences of some of the components of MS were also observed.Low HDL-c was 61.2% in this study, close to the value obtained in Limache (66.7%), but higher than the prevalence obtained in RP (42.2%).Raised blood pressure was observed in 15.5% in Limache and 23.1% in RP whereas in this study a higher prevalence was estimated, 27.1%.Abdominal obesity was observed in 19.3% in Limache and 12.7% in RP whereas in this study its prevalence was 13.7%.It is noteworthy that in this study the prevalence of smoking was much lower (4.3%)than in Ribeirão Preto (17.2%) or Limache (56.4%).In contrast, the prevalence of a sedentary lifestyle was far greater in this study (69.6%) than in Ribeirão Preto (33.2%) or Limache (38.3%).
Bustos et al. 13 , using the NCEP ATPIII definition, showed a 10.1% prevalence of MS in Limache and 7.6% in Ribeirão Preto.In this study, the prevalence of MS by the same definition was 14.9%, higher than the values observed in these studies.These differences could be explained by the time of completion of the studies, as the data in Limache and Ribeirão Preto are from 2002/2004 while data in this study are from 2011/2012.In addition, these differences might be due to the fact that Ribeirão Preto and Limache samples were drawn from the entire population whereas this study sample was drawn from the population of university students that tend to be better off.Yet paradoxically in this study the prevalence of RI was 7.3%, which is much lower than that observed in Ribeirão Preto in 2002/2004 (13.9%) and Limache (34.5%).Differences in the cutoff points used (IR > 2.7 in this study and IR > 2.53 in the other studies) would not explain such large differences.In Ribeirão Preto, the prevalence of MS was higher among men, which agrees with data from this study.However, the prevalence of IR in Ribeirão Preto was also higher among men, while in this study there was no difference by sex.
In a study of 2,130 adults aged greater than 18 years old in Brasilia, a 32% prevalence of MS by the JIS criterion 24 was observed.The prevalence of MS in the previous study in the youngest age group (18-24 years) was 24.8%, higher than that found in this study by the same criterion (20.5%).In Brasilia, although there was no difference in prevalence between men and women, among women, however, higher education was a protective factor for metabolic syndrome 25 .We found that in this population of high education students, women had a lower prevalence of MS than men.
In another population-based study in Vitória with adults aged 25 to 64 years in 1999 and 2000, the prevalence of MS by the NCEP definition was 29.8%, with no difference between sexes 26 .The prevalence of MS by the NCEP criterion among 25-34 year olds (15.8%) in Vitória was very similar to that observed in this study for the younger age group (14.9%).Another study by Marquezine et al. 27 with the urban population of Vitoria, found a lower prevalence of MS by the NCEP in the young age group, 12.3% for the age group 25-35 years, compared to this study.In this study, the prevalence of MS by the NCEP was higher for males (22.1%) than for females (10.7%, p = 0.019), which differs from most studies that did not find a significant difference in the prevalence of MS according to sex.
The high prevalence of MS among male college students in the sample studied in São Luís, may be explained by the fact that the student population is composed of individuals of higher income and education.The worse cardiovascular risk profile observed in men seems to be a portrait of the most advantaged social groups.At the University of Kansas, in the United States, men were more obese, more hypertriglyceridemic and hypertensive than women 28 , in accordance with findings from this study.This study also showed that in this population of university students, composed predominantly of middle and high-income individuals, paradoxically, men were more obese than women, despite being more physically active and having a lower percentage of high fat intake.Part of the explanation for this apparent paradox could be a higher intake of carbohydrates by men, but unfortunately, we did not obtain this information.As this is a cross-sectional study, another possible explanation is that men being more obese have so practiced more physical activities and ate less fat, trying to lose weight.The association between low fat intake and a less sedentary lifestyle with worse cardiovascular risk profile among men would then be explained by reverse  31 .The big difference in the prevalence of hypertension among men (46.3%) and women (15.3%) detected in this study, was probably due to the fact that most risk factors for cardiovascular disease are more prevalent in males than females.This difference was also found in other studies 10,30,31 .
We found a very high prevalence of low HDL-C (61.2%).High prevalence of low HDL-C is associated with increased waist circumference and a sedentary life style 32 , and both were highly prevalent in the current study (36.0% and 69.6%, respectively).In some studies, a high prevalence of low HDL-C was also observed, although smaller than that found in the present stud: 44.1% in a A higher prevalence of cardiovascular risk factors found among students from private compared to those who attend public institutions can be justified by the higher income of students from private universities.National studies show a higher prevalence of overweight (p = 0.004) 34 and increased waist circumference (p = 0.008) 35 in high-income individuals.
Among the strengths of this study we outline: it is a population-based sample of young adults, age at which few population studies have been conducted in Brazil.The percentage of losses was satisfactory (24.1%).Participation rate was higher for students from public (78%) than private (72%) institutions.However, the participation rate for laboratory examinations was low, 61%, being higher for public (78%) than private (49%) universities.Fewer university students with higher family income, higher educational level of household head, private institutions, from catholic backgrounds, who had no religion or who had health insurance attended blood collection.Due to the smaller percentage of attendance for laboratory tests, prevalences were weighted by the inverse probability of selection of each individual, taking non-response into account.It has been shown that in the presence of losses, inverse probability weighting is a method that reduces selection bias 36 .Prevalence of MS according to JIS was 18.1% and increased to 20.5% after inverse probability weighting taking non-response into account was used, which shows that this method was able to attenuate underestimation of MS in the population.Even with this correction it is still possible that the prevalence of MS is underestimated in this study if not all predictors of participation in laboratorial exams were included in the logistic regression model.

collaborations
JB Barbosa worked in the study design, analysis and interpretation of data, and drafted the article; AM Santos worked on the analysis and interpretation of data; MM Barbosa, JM Fonseca, MCL Barbosa and MM Barbosa collaborated in the study design and helped in data interpretation; CA Carvalho and PCA Fonseca worked in the study design and data analysis; EG Bogéa worked in data analysis and interpretation, and wrote the final version of the manuscript; AAM Silva conceived the study and helped in data analysis and interpretation.All authors reviewed and approved the final version.conclusion High prevalences of metabolic syndrome (20.5% according to the JIS definition), insulin resistance (7.3%) and other cardiovascular risk factors were found in this young population.This suggests that the burden of these diseases in the future will be increased, especially among men and students from private universities.

acknowledgments
We thanked the Maranhão State Research Foundation (FAPEMA in the Portuguese acronym), UDI Hospital and Laboratório Cedro for providing financial and logistic support.

table 1 .
Characteristics of the university students of São Luís, Maranhão, Brazil, 2011-12.Totals do not add to 968 because of missing values ProUni: University for All Program. a
a P value calculated by the Student's t test when variables presented normal distribution or the Mann-Whitney test otherwise; b Variables with non-normal distribution: values represent median (25th percentile -75th percentile); c Variables with normal distribution: values represent mean (standard deviation); d Body fat percentage estimated by bipolar bioelectrical impedance.HDL: High Density Lipoprotein; HOMA -IR: Homeostasis Model Assessment -insulin resistance.
a P value calculated by the chi-square test; b Physical activity evaluated and classified according to guidelines from the International Physical Activity Questionnaire; c Block score -high fat consumption -score > 27; d High waist-to-hip ratio ≥ 0.90 for males and ≥ 0.85 for females; e High waist-to-height ratio ≥ 0.53 for males and ≥ 0.52 for females; f Body fat percentage measured by bipolar bioelectrical impedance -obesity ≥ 25% for men and ≥ 32% for women, malnutrition < 6% for men and < 8.9% for women.

table 4 .
Metabolic syndrome and its components, and insulin resistance among university students according to sex, São Luís,Maranhão, Brazil, 2011-12.High waist circumference according to the NCP criterion ≥ 102 cm for males and ≥ 88 cm for females; c High waist circumference according to the IDF criterion ≥ 90 cm for males and ≥ 80 cm for females; d Hypertension -systolic blood pressure ≥ 130 mmHg and/or diastolic blood pressure ≥ 85 mmHg; e Hyperglycemia ≥100 mg/dl; f Hypertriglyceridemia ≥ 150; g Low HDL-cholesterol < 40 mg/dl for men and < 50 mg/dl for women; h HOMA -IR: Homeostasis Model Assessment -insulin resistance > 2.7.HDL: High Density Lipoprotein; NCEP: National Cholesterol Education Program; IDF: International Diabetes Federation; JIS: Joint Interim Statement.
a P value calculated by the chi-square test; b

table 5 .
Lifestyle and anthropometric variables among university students attending public or private institutions, São Luís, Maranhão, Brazil, 2011-12.P value calculated by the chi-square test; b Physical activity evaluated and classified according to guidelines from the International Physical Activity Questionnaire; c Block score -high fat consumption -score > 27; d High ratio ≥ 0.90 for males and ≥ 0.85 for females; e High waist-to-height ratio ≥ 0.53 for males and ≥ 0.52 for females; f Body fat percentage measured by bipolar bioelectrical impedance -obesity ≥ 25% for men and ≥ 32% for women, malnutrition < 6% for men and < 8.9% for women. a

table 6 .
Metabolic syndrome and its components, insulin resistance among university students attending public or private institutions, São Luís,Maranhão, Brazil, 2011-12.High waist circumference according to the NCP criterion ≥ 102 cm for males and ≥ 88 cm for females; c High waist circumference according to the IDF criterion ≥ 90 cm for males and ≥ 80 cm for females; d Hypertension -systolic blood pressure ≥ 130 mmHg and/or diastolic blood pressure ≥ 85 mmHg; e Hyperglycemia ≥100 mg/dl; f Hypertriglyceridemia ≥ 150; g Low HDL-cholesterol < 40 mg/dl for men and < 50 mg/dl for women; h HOMA -IR: Homeostasis Model Assessment -insulin resistance >2.7.HDL: High Density Lipoprotein; NCEP: National Cholesterol Education Program; IDF: International Diabetes Federation; JIS: Joint Interim Statement.
a P value calculated by the chi-square test; b

table 7 .
Comparison of some characteristics between those who did or did not attend blood collection, São Luís, Maranhão, Brazil, 2011-12.P value calculated by the chi-square test. a