This is an observational, descriptive, and cross-sectional study. The participants were elderly (above 60-years old), of both genders, and were community-dwelling in Teresina, Piauí-Brazil. Subjects were selected using data from a Brazilian Ministry Health Program (Health Family Strategy). The sample size was calculated from the standard deviation estimative of 28.5 ng/mL of 25(OH)-D [11] , with a confidence level of 95% and sample error of 5.6 ng/mL. We excluded subjects that were bedridden, those that used a wheelchair, those with hearing or talking impairments, those who were taking vitamin D or calcium supplements and those with diseases related to vitamin D metabolism (for instance, kidney or liver diseases). The guidelines for human studies were followed according to the Declaration of Helsinki. The volunteers elected to the study signed their informed consent, and the project was approved by the Ethics Committee of the Federal University of Piauí, protocol number 0386.0.045.000-10.

Data were obtained during a house interview, adopting a questionnaire that had been pre-tested in a previous pilot experiment. Data were collected from February to June, 2011. The anthropometric measurements were: body weight (Plenna scale, 150 Kg), knee height-measured with a caliper consisting of an adjustable measuring stick with a blade attached to each end, at a 90˚ angle [20] [21] —and waist circumference (WC), measured with an inelastic fiberglass tape. From the measurements, we estimated the height [20] , and calculated the body mass index [BMI = weight/(height)²]. To classify the nutritional status according to BMI, we adopted the recommendation of the Pan-American Health Organization; normality when BMI was between 23 - 28 kg/m², overweight when 28 kg/m² ≤ BMI < 30 kg/m² and obesity when BMI ≥ 30 Kg/m² [22] . To identify cardiovascular risk from the waist circumference, we adopted standards proposed by the WHO [23] , where waist circumference ≥ 102 cm for men and ≥ 88 cm for women constituted cardiovascular risk.

Blood pressure was measured following the VI Brazilian Guidelines for Hypertension (BD sphygmomanometer). According to these guidelines, hypertension was defined as systolic blood pressure (SBP) ≥ 140 mm/hg and/or diastolic blood pressure (DBP) ≥ 90 mm/Hg. The values were obtained by calculating the average of two measurements [24] . For participants with diabetes, hypertension was defined when SBP was ≥ 130 and/or DBP 80 ≥ mmHg. In addition, participants who referred to taking medicines for hypertension were also considered as having high blood pressure [25] .

For blood sample collection, the participants were required to attend the Health Unit on a scheduled day, during the morning, after fasting and rest of between 8 - 12 hours. In case of any difficulty, the researcher went to the participant’s house to carry out blood withdrawal. The blood was centrifuged and the plasma was frozen until analysis. Biochemical analyses were performed at the Hospital Med Imagem, Teresina, Piauí, Brazil. PTH and 25(OH)-D levels were analyzed by chemiluminescence method (IMMULITE 200 and Diasorin LIAISON^TM, USA, respectively). The reference range for PTH was 12 to 72 pg/mL, and the classification of vitamin D levels followed that proposed by Holick et al. [26] : ≤ 20 ng/mL-vitamin D deficiency; 21 to 29.9 ng/mL-insufficient level; ≥ 30 ng/mL to 100 ng/mL-adequate levels. Plasma lipids were analyzed by automation (ViteK Systems Clinline 150-Biomerieux). HDL-cholesterol and triacylglycerol were measured by enzymatic colorimetry, and LDL-cholesterol was calculated according to the Friedwald equation [27] . We adopted reference values according to the IV Brazilian Guidelines on dyslipidemia and arterosclerosis prevention [28] . Fasting blood glucose was measured by enzymatic colorimetric method, and the reference values followed those of the Brazilian Society of Diabetes [29] . Participants with blood glucose values ≥ 126 mg/dL and those with normal values, but under pharmacological treatment, were considered as diabetic.

Numeric data were presented as mean and standard deviation. The categorical variables were presented as percentages. The Kolmogorov-Smirnov test revealed that none of the investigated variables presented normal distribution. Therefore, we adopted non-parametric tests (Mann-Whitney test and Kruskal-Wallis test). The association between categorical variables was investigated by the Chi-square test (c²), and cases where more than 25% of the cells presented n < 5 we adopted the Fisher exact-test. Some data were correlated by Spearman test. We considered a p-value < 0.05 as significant for all analyses and the analyses were performed with the Statistical Package for Social Sciences (SPSS^®), version 18.0.

Table 1 depicts the variables investigated and Table 2 presents the prevalence of cardiovascular risk, for both genders. With regard to demographic data, the majority of the participants declared living with a partner; most of them reported low schooling and low income.

Both men and women presented a low prevalence of total obesity, as evaluated by BMI values. However, abdominal obesity, evaluated by WC, suggested that the majority of women presented CVD risk, but not the majority of men. Both genders presented an elevated prevalence of high blood pressure.

Biochemical analysis suggested low levels of 25(OH)-D for both genders, but women presented lower values than men. PTH values were between normality ranges for the majority of both men and women. High fasting glucose was relatively low in prevalence both in men and women, although the mean values were higher in men. The prevalence of high total cholesterol and triacylglycerol was elevated in both genders, without any difference between genders. In turn, the prevalence of both high LDL-cholesterol and low HDL-cholesterol was lower than other markers of blood lipids, although women showed a higher prevalence than men. Alcohol intake and smoking were referred by a small percentage of the participants; men reported these habits more than women. In addition, a sedentary lifestyle was referred by the majority of participants, of both genders.

In order to turn our study broader, we investigated cardiovascular risk by taking into account other independent variables. With regard to schooling, total obesity was higher in the subjects with higher schooling level. In contrast, a sedentary lifestyle was more common in subjects with lower schooling (illiterate = 7.2% and 81.3% for total obesity and sedentary lifestyle, respectively; fundamental level = 12.4% and 73.3% respectively; intermediate level and above: 16.9% and 55.9% respectively; p < 0.05 for all the comparisons). Biochemical variables and blood pressure values did not present any significant association with regard to schooling. Considering marital status, high blood pressure, abdominal obesity and low HDL-cholesterol were more prevalent in the

^*Category variables were compared by Chi-squared test (c²), and numeric variables were compared by Mann-Whitney Test. ¹Complete or incomplete; ²MS = Minimal salary; ³BMI = body mass index; ⁴SBP = systolic blood pressure; ⁵DBP = diastolic blood pressure.

participants living without a partner (with partner = 65.3%, 56.9% and 23.2%, respectively; p < 0.05 for all comparisons; without partner = 81.3%, 79.1% and 45.2%, respectively). In turn, alcohol intake was higher in participants living with a partner (27.1%, versus 14.9%; p < 0.05). There were no differences in any of the risk factors when stratified for income. When grouping the participants according to age intervals (60 - 69, 70 - 79 and above 80 years), the only variable that was significantly different in the groups was alcohol intake (29.5%, 15.6% and 14.8% respectively, p < 0.05).

Table 3 shows the correlation between vitamin D and variables related to cardiovascular risk, and Table 4 shows the correlation between PTH and the same variables. We did not observe any significant correlation of vitamin D levels and any of the variables. With regard to PTH, only DBP presented significant correlation with this factor.