ABC | Volume 112, Nº3, March 2019

Original Article Sousa et al Prevalence of hypertension in older adults Arq Bras Cardiol. 2019; 112(3):271-278 Methods This is a cross-sectional population-based study, carried out through a household survey and randomized cluster sampling, from the matrix project “ Situação de saúde da população idosa do município de Goiânia-GO ” (Health Status of the Elderly Population of the Municipality of Goiânia-GO), linked to the Rede de Vigilância à Saúde do Idoso (REVISI) (Health Surveillance Network of the Elderly) in the State of Goiás. The methodological and sample calculation details were described in previous publications. 13,14 This study was developed by Universidade Federal de Goiás, the Municipal Health Secretariat of Goiânia and State Health Secretariat of the state of Goiás, through REVISI, after being approved by the Research Ethics Committee of Universidade Federal de Goiás (Protocol number 050/2009), in agreement with the Declaration of Helsinki. An epidemiological survey was carried out, with the participation of people aged 60 and over, living in their homes and in the urban area of Goiânia. The sample was calculated considering the elderly population as 7% of the total 1,249,645 inhabitants, based on the year 2007, 15 an estimated frequency of 30% (lowest expected frequency among the variables investigated in the matrix project), a 95% confidence interval (CI), a 5% significance level, and 5% absolute precision. To the calculated sample (n = 823), 11% were added to compensate for losses, and 934 elderly were assessed. Of the total of 934 questionnaires, 22 were excluded due to data inconsistency, and the final sample consisted of 912 elderly individuals. The study area was defined based on census sectors (CS). The sample units were the households and the elderly in elementary observation units. Firstly, the CS were identified using the Basic Urban Digital Map of Goiânia as the basic layer. The sampling process was based on the maps of blocks and allotments of the selected regions and was carried out in multiple stages starting from the identification of CS defined by the Instituto Brasileiro de Geografia e Estatística (IBGE - Brazilian Institute of Geography and Statistics). 16 A total of 56 CS were selected, with an estimate of reaching, on average, 17 elderly individuals in each CS. The data were collected from residents who were at home at the time of the interviewer’s visit and who accepted to participate in the study by signing the Free and Informed Consent Form. If, during data collection, two consecutive households with elderly residents were identified, the second house was excluded to minimize the conglomerate and neighborhood effect. The following inclusion criteria were considered in the study: age older than 60 years and being a resident of the household. Elderly individuals who were at the household at the time of the interview but were not residents or were unable to answer for any reason (dementia, unconsciousness) were excluded. In those cases, that household was disregarded, and the next house was considered. The interviews were carried out by researchers properly trained to apply the study forms and also for the standardization of the procedures to be performed in data collection. The interviews were carried out from November 2009 to April 2010, considering the baseline for the Rede de Vigilância à Saúde do Idoso (Health Surveillance Network of the Elderly) in the capital city. Further details on the method can be verified in a previous publication. 17 At the time of data collection, information were obtained on age, gender, socioeconomic status (level of schooling, marital status and family income), modifiable risk factors (physical activity, smoking, alcohol consumption) and information on AH treatment. Blood pressure (BP) levels were also measured. BP was measured using an OMRON automatic device, model HEM-705CP, following the protocol of the Brazilian Guidelines. 18 Three measurements were performed in the same arm with the person in the sitting position, following a 3-to-5 minute interval, using the last two measurements for the calculation of the mean value, providing the difference between them was not greater than 4 mmHg. This was done to reduce data dispersion. It is worth noting that appropriate cuff sizes were used according to the arm circumference, using adequate sizes (standard, obese, pediatric) that covered two-thirds of arm extension. 18 To identify AH prevalence, the elderly were considered hypertensive if they, during data collection, had systolic pressure values ≥ 140 mmHg or diastolic pressure ≥ 90 mmHg, or if they reported regular use of antihypertensive drugs, regardless of the BP value at the time of the interview. 18 All patients who reported antihypertensive medication use at the time of data collection and who were able to show the prescription, or the medication boxes to be verified, were considered as undergoing treatment for AH. The individual was considered to have controlled BP when he/she reported AH treatment and the mean BP value was lower than 140/90 mmHg. Smoking status was classified into three groups: ex-smokers, regardless of the time since they had stopped smoking, non-smokers, for those who never smoked, and smokers. Alcohol consumption was identified according to the elderly individual’s response into two groups: those who reported consuming alcohol, even occasionally, and those who reported not consuming it at all. Individuals who reported regular physical activity (three or more times a week) were classified as non-sedentary, and those who practiced physical activity less than three times a week or did not practice any physical activity were classified as sedentary. Statistical analysis The quantitative variables were shown with their means and medians, standard deviations and 95%CI; the categorical variables were shown, according to their frequencies, as absolute numbers and percentages. The analysis of the normal distribution of data was performed using the Kolmogorov-Smirnov test. The software SPSS-IBM version 23 was used to analyze the data, and the odds ratios, AH prevalence, treatment and control rates were calculated, with 95%CI. The chi‑square test was used to analyze the association between AH and categorical variables, and the Mann Whitney-U test of independent samples was used to analyze the association between non-parametric, continuous quantitative variables. 272