Cardiovascular Risk Factors and their Relationship with Educational Level in a University Population

1Faculdade Santa Terezinha – Setor de Fisioterapia – São Luís, MA – Brazil 2Universidade Federal do Maranhão – Departamento de Pós-graduação em Saúde Coletiva – São Luís, MA – Brazil 3Universidade Federal do Maranhão – Departamento de Medicina I – São Luís, MA – Brazil 4Universidade Federal do Maranhão – Departamento de Medicina III – São Luís, MA – Brazil 5Hospital de Messejana – Fortaleza, CE – Brazil 6Universidade Federal do Maranhão – Hospital Universitário Presidente Dutra – São Luís, MA – Brazil


Introduction
Most studies associate the lowest level of education to the highest prevalence of cardiovascular risk factors (CVRF), as shown by Barros et al. 1 and Santos et al. 2 Vaydia et al. 3 , however, showed that even among individuals with some knowledge about cardiovascular health, there is a gap between knowledge and practice of healthy habits.
In this context, studies are controversial regarding the influence of the level of education on the prevalence of cardiovascular risk factors: for some CVRFs, level of education appears to be protective, but for others it seems to have no influence at all [4][5][6][7] .
Aiming to determine the prevalence of cardiovascular risk factors, as well as the influence of education and teaching practice on their presence, this research was conducted on a group of federal public university servants.

Methods
A cross-sectional, analytical study used the database of a cohort referred to as "cardiovascular risk factors in public servants at Universidade Federal do Maranhão (UFMA)".Study data were collected from March 2009 to October 2010, involving active public servants having employment relationship with UFMA, in the city of São Luís, Brazil.
The study protocol was submitted for approval by the Research Ethics Committee of Hospital Universitário Presidente Dutra / UFMA, under no.003835/2008-80.
The list of physically active servants of the University (n=2 238) was provided by its Human Resources Department.The sample calculation considered a confidence index of 95%, α=5%, sampling error=2%, and CVRF prevalence of 50%.An estimated sample of 350 public servants was obtained (plus 10% for possible losses).
All randomly selected public servants, who were physically active, of both genders, aged over 18 and who agreed to participate in the research were included.All public servants who were in retirement process, refused to participate in the research and were illiterate were excluded.
A random sample of 350 public servants was located per Center and Department to which they were assigned and position held.All participants signed the Informed Consent Form.
Data was collected by healthcare academic students of the university, supervised and previously trained.
Anthropometric measures such as weight, height and waist circumference were obtained by trained personnel.All the participants in the original project answered to the CVRF, SH presence or DM questionnaire, and provided treatment evidence in the form of presentation of medication, medicine packaging or prescription.These diseases have been proven to be controlled by means of laboratory tests and clinical evaluation.Participants were weighed on a Plenna® digital electronic scale, Acqua model, accurate to 100 g (Guangzhou, China).Height was measured in centimeters, using a 0.1 cm-resolution, portable stadiometer.All measurements were taken twice in a row and the mean value was used in the analyses.
Blood pressure was measured using a mercury sphygmomanometer, with a blood pressure cuff of the same size, adjusted to the arm circumference.The procedure was carried out three times, at 10-15 minute intervals, with the mean value between the last two measurements considered, pursuant to the methodology used for epidemiological studies.The adopted reference values followed the recommendations of the VI Diretriz Brasileira de Hipertensão Arterial (VI Brazilian Guidelines on Hypertension) 9 : high blood pressure levels ≥90 mmHg diastolic blood pressure and ≥140 mmHg systolic blood pressure.Individuals who had blood pressure below these levels, but reported to have been using antihypertensive drugs, were also considered hypertensive.
Participants were instructed to fast for 8- For the level of physical activity and sedentary lifestyle to be diagnosed, the participants who have reported having done regular, mild or moderate-intensity physical activities, for less than 150 minutes or vigorous-intensity physical activity for at least 75 minutes a week, according to the World Health Organization 12 and the World Heart Federation 13 .
Smoking was classified as: never (those who reported having never smoked and those who had stopped smoking for six months or more), ex-smokers (those who reported not to smoke currently, but had already smoked for six months or more) and smokers (those who reported to smoke, regardless of how much or how long they have been smoking).
For overweight and central obesity to be diagnosed, the body mass index (BMI), which is a function of body mass and body height (body weight (kg) divided by the square of the body height (cm 2 )), was calculated.
Waist circumference (WC) was measured pursuant to the NCEP III 11 with flexible but inelastic measuring tape.
The cutoff points adopted were those according to the risk level for cardiovascular disease: increased risk for women (WC> 80 cm) and men (WC > 94 cm) and highly increased risk for women (WC> 88 cm) and men (WC> 102 cm).
Hip circumference (HC) was measured using a flexible but inelastic measuring tape.
Nine changeable variables were selected, which are listed below, with their abnormality criteria.The cutoff points were based on the World Health Organization, NCEP ATP III 11 and Brazilian Society of Cardiology 9 .
• Overweight: BMI (weight/height 2 ) ≥25 kg/m² and <30 kg/m²; • Obesity: BMI (weight/height 2 ) ≥30 kg/m²; • Central obesity: waist circumference ≥88 cm for women and ≥102 cm for men, regardless of the presence of widespread obesity; • High total cholesterol: ≥240 mg/dL, according to the enzymatic Trinder method; • HDL-c <40 mg/dL for men, and <50 mg/dL for women, according to the Labtest method; • Triglycerides: ≥150 mg/dL, according to the modified Soloni method; Other independent variables included were not considered changing factors: age, gender, race (biological variables).Race was defined in this study using skin color as reference (white vs. nonwhite), according to the official nomenclature used in demographic censuses.
Data were analyzed using SPSS, version 21.Qualitative variables were expressed as frequencies; quantitative variables as means and standard deviations.
The t-Student test was used for independent samples to compare the means of quantitative variables between the levels of education.CVRF association with the groups was verified using the Chi-Square Test for independent samples.Both tests were performed with a 95% confidence interval; significance level α=5%; sampling error=2%.
A multivariate logistic regression analysis was performed.The variables with p < 0.05 in the univariate analyzes were included in the models and were considered statistically significant.

Results
From the 350 randomly selected public servants, 31 were excluded for not meeting the inclusion criteria.A representative sample of 319 public servants, a percentage found within the loss patterns of 10% calculated for the study, remained.Table 1 shows the sociodemographic characteristics of the study sample.Individuals without higher education level accounted for 43.89%, while those with higher education accounted for 56.11%.Levels of education were evenly distributed by gender (p=0.735),age (p=0.676) and marital status (p=0.365), with higher prevalence of non-white, born in the capital city, assigned to the university campus.Teachers were mostly male, aged 50-59, non-white, married, born in São Luís and assigned to the university campus.
When CVRF prevalence of both groups -with and without higher education -is compared, the group with higher education has shown to be less sedentary (p=0.003) and to have a lower prevalence of DM (p=0.033).Education levels were evenly distributed among overweight, obesity, systemic hypertension, dyslipidemia, and smoking (Table 2).With regard to teaching practice, the proportion of cardiovascular risk factors was similarly distributed between teacher and non-teacher groups (Table 3).
Table 4 shows the blood pressure levels, anthropometric indices, lipid profile and glucose level according to the level of education.Except for the glucose level, which was higher in the group without higher education (98.67±34.08mg/dL vs.92.63±13.64mg/dL; X²=0.031); the other blood pressure, anthropometric and lipid variables were similar in both groups.
The binary logistic regression results are described in Table 5.

Discussion
The association of education level with CVRF in university environments has been assessed and appears not to be defined.In general, the best level of education is associated with a lower negative perception of health 14 , with individuals with higher education being expected to have more knowledge about health and healthier habits 15 .Thus, the results of this work can partly be attributed to the existence of a gap in the relationship between knowledge of cardiovascular health, attitude and healthy practice/behaviors, a hypothesis already formulated by Vaidya et al. 3 In this study, DM prevalence result was 9.4%, a value higher than the global prevalence values and estimates, considering individuals of the same age, as shown by Shaw et al. 16 and Guariguata et al. 17 .When compared to Brazil, this study showed values greater than those of the recent population-based research 18 , which shows an increased prevalence of DM between 2006 and 2013, from 5.5% to 6.8%.
Studies conducted in other universities have shown different DM prevalence values, with higher rates found in the Brazilian southeast region 19 and in the Longitudinal Study of Adult Health (ELSA) 20 .In the latter, the value is twice as high as that found in this study.
The differences of the studies can be explained by methodological issues (randomization, telephone survey, self-report, use of laboratory tests and the different glucose cutoff points), which may influence on the prevalence of DM.Another important point lies in the differences arising from the development level of countries.In this sense, Reddy 21 notes that in developing countries, the industrialization process entails greater exposure to CVRF due to changes in lifestyle, such as in dietary habits and physical activity.
Regarding the association between DM prevalence and level of education, other studies have shown conflicting results, such as that conducted in nine European countries, which notes that there is no difference between glucose levels among individuals with different education levels 22 , as well as in a similar research conducted in Africa, in which the distribution of DM was similar among the different levels of education 5 .However, studies carried out in Spain 23,24 show higher prevalence of DM in people with low education levels.In a population-based survey, carried out in 2012, the DM prevalence between the lowest and the highest education level ranged from 12.1% to 3.8% 4 .
This association of DM with the level of education is complex and has significant consequences, such as those shown in the cohort study with 85,867 individuals in the United States, which showed that adults with DM and low education levels have mortality rate 503 times higher per 10,000 inhabitants than those patients with higher levels of DM 26 .
It is worth of notice that the teachers participating in this study have shown DM prevalence of 13.1%, approximately ¼ than that shown in a similar study in the Brazilian southern region 27 .When prevalence between teachers and non-teachers was compared, no significant difference was found between both groups.However, when comparing both groups regarding education level, blood glucose levels have shown to be higher in individuals with lower education level, a result similar to that given in the ELSA study 20 and in an African study by Capingana et al. 5 Different results are found in the study by Achidi et al. 7 , where teachers had higher glucose levels than non-teachers.This contradictory result is suggested to be explained based on the fact that, in this study, about 60% of teachers are non-white and, according to Santos et al. 2 , blacks and mulattos do not differ from white as to smoking habits, sedentary lifestyle and the presence of hypertension, a fact that may have influenced the results found.
The prevalence found for sedentary lifestyle was 72.1%, a value roughly equivalent to double when compared to the highest rate of sedentary lifestyle found by Lloyd-Sherlock et al. 28 , which ranged from 24.9% to 37.6%.Compared to Welmer et al. 29 , the value found is four times as high as when similar age groups are observed.
A Brazilian study  18 .
The inverse association between level of education and sedentary lifestyle observed in this study has also been described in European studies 23,30 .In this context, Morales-Asencio et al. 23 emphasize how patterns of behavior relate to socioeconomic and cultural factors.
A study by Petroski 27 has found prevalence of sedentary lifestyle among college teachers 25% lower than that of this study.Reis 14 and a study at a university center in Russia 31 , however, have shown higher values.A significant amount (42.6%) of Russian teachers do not practice any physical activity and explain that this fact is due to their lack of time, activities or studies after working hours, as well as the lack of appropriate conditions for physical activities.Another study adds other barriers such as lack of self-discipline, lack of equipment and lack of a person to keep company during the activities 32 .As in this research, a study conducted by a Brazilian university 33 was also contradictory to Petroski's findings 27 .
It is worth highlighting that the obesity rate found in the participants was lower than that found in South America and Brazil (2007-2013) 34 , while it is in line with studies conducted in São Luís, in 2009 35 38 , and this difference is due to the fact that participants had not been randomly selected.
Contrary to DM and sedentary lifestyle, obesity was evenly distributed according to the education level, differing from the African study 5 , from that carried out by Santos et al. 2 , and from VIGITEL Research 18 , which state that a lower prevalence of overweight and obesity is typical in a population with a higher education level.
One reason for the divergent association between level of education and obesity in Brazil was found in the Pro-Health study conducted by Fonseca et al. 33 , in a university center, which grounded this association on cyclical changes in weight.The study revealed that both continuous and cyclical changes in weight throughout life were associated with the level of parental education and that of the individual themselves, with significant gender-related differences, whereas for men, this association was not confirmed.Women with lower level of education have shown a chance higher than 94% of reporting a cyclical increase in body weight.
By contrast, a study similar to this research, carried out in Africa 7 , indicates no association between body mass index and education level, according to its findings.
Being teachers with graduate degrees (master's and doctoral) was not associated with lower prevalence of obesity when compared to the group of non-teachers.This result contradicts the findings of Hayes et al. 15 , whereby people with higher education level have more access to health information, better social support, more awareness of the importance of staying healthy and more timely access to healthcare services.
The SH, smoking and overweight values shown by the participants in this study are higher and diverge from those reported by Malta et al. 39 According to these authors, the CVRF are mostly common in less educated adults, as in this study the levels of education were evenly distributed.However, a European study shows that smoking was not controlled by the lower level of education, a result similar to this study 40 .
The high prevalence of overweight in the population studied is in line with recent research findings that averaged between 48.5% and 54.7% 18,38 .As for the prevalence of dyslipidemia found, this is more than twice as high as that found in a recent Brazilian study 18 .
That research was conducted with a random sample of the population at a public university, with clinical and laboratory tests being conducted to determine the presence of cardiovascular risk factors in the individuals studied.However, since the study was cross-sectional, it was not possible to establish a causal relationship in these associations.The limitations to this study lie in the failure to check the consumption of alcohol and the absence of individuals with an education level of less than five years.

Conclusion
This study has shown that individuals with higher levels of education had lower prevalence of DM and sedentary lifestyle.

Potential Conflicts of Interest
No relevant conflicts of interest.

Sources of Funding
This study was partially funded by a Master's scholarship awarded by CAPES -Coordenação de Aperfeiçoamento de Pessoal de Nível Superior.

Academic Association
This manuscript is part of the Master's dissertation of Rute Pires Costa from Universidade Federal do Maranhão.
ABBREVIATIONS AND ACRONYMS

Table 1 Sociodemographic characteristics of the study population according to the level of education and teaching practice
*Group without higher education level (elementary and middle school); **p < 0.05; CI = 95%; Chi-square test for independent samples.HUUFMA -Hospital Universitário da Universidade Federal do Maranhão

Table 2 Cardiovascular risk factors in the study population, according to education level
SH -systemic hypertension; * p<0.05;CI=95%; Chi-square test for independent samples

Table 3 Cardiovascular risk factors in the study population, according to teaching practice Variables Non-teachers Teachers p-value n (%) n (%)
p<0.05; CI=95%; Chi-square test for independent samples SH -systemic hypertension

Table 4 Blood pressure levels, anthropometric indices, lipid profile and glucose level of the study population, according to the level of education
*p<0.05; CI=95%; t-Student test for independent samples *Values expressed as mean±standard deviation.SBP -systolic blood pressure; DBP -diastolic blood pressure; BMI -body mass index; WHR -waist-to-hip ratio

Table 5 Independent variables associated with the level of education in the study population Variables Odds ratio CI95% p-value
2 , with 2,531 individuals of different races, from 72 cities, shows high sedentary lifestyle values, ranging from 76.6% to 81.83% and, therefore, similar to this research.A lower value (16.2%) is found in a Brazilian population study conducted in 2013 36d 201334.Interestingly, this capital has a lower countrywide prevalence of obesity, with Cuiabá and Rio de Janeiro showing the highest prevalence values.Studies in universities, such as those by Achidi et al.7, Scott et al.36and Duncan et al. 37 indicate higher values.By contrast, a Southeastern Brazil study shows a value equivalent to half that observed in this research