Big data e tele-eletrocardiografia: avaliação da mortalidade em pacientes portadores de fibrilação atrial

• Ano de defesa: 2018
• Autor(a) principal: Gabriela Miana de Mattos Paixao
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Coorte Eletrônica; Mortalidade; Fibrilação Atrial; Eletrocardiograma; Telemedicina; Eletrocardiografia; Medicina
• Link de acesso: http://hdl.handle.net/1843/BUOS-B95HZZ

Introduction: Telehealth system is an important tool to improve access and quality to health assistance, especially in poor regions. The use of digital ECGs in primary care population provides clinical and epidemiological diagnoses with implications in healthcare policy. Large electrocardiogram (ECG) databases, linked to mortality or hospitalization data, can be very useful in determining the prognostic value of established and new ECG markers. Atrial fibrillation (AF) is a public health problem with increasing prevalence as the population ages, associated with cardiovascular mortality and morbidity. Objective: Evaluate the association between the presence of AF with overall and cardiovascular mortality in a large electronic cohort of primary care patients of Minas Gerais. Methods: This is an observational retrospective study. Patients over 16 years old who performed digital electrocardiograms by Telehealth Network of Minas Gerais from 2010 to 2017 were assessed. A probabilistic linkage between data from the national mortality information system and our ECG database was made. Clinical data were self-reported, and ECGs were interpreted by a team of trained cardiologists and automatic software (Glasgow and Minnesota).The diagnosis of AF was considered if there was concordance between the cardiologist´s report and one of the automatic systems. In cases of disagreement, ECGs were reviewed manually. Only the first ECG made was analyzed. To assess the relation between AF and mortality, Cox regression was used, adjusted by age, sex and clinical conditions. Results: From a dataset of 1,773,689 patients, 1,550,959 were included, mean age 51.66 years; 40.24% male. There were 3.35% deaths for all causes in 3.69 years of mean follow up. The prevalence of AF was 1.38%. In univariate analysis, AF was a risk factor for death from all causes (HR 5.93, 95%CI 5.746.13). After adjustment for age, sex and comorbidities, AF remained an independent risk factor for all-cause mortality (HR 2.46, 95%CI 2.38-2.55). AF was also a predictor of risk for cardiovascular mortality after adjustment for age, sex and clinical conditions (HR 2.28, 95%CI 2.04-2.55). Female patients with AF had higher risk of overall and cardiovascular mortality compared with male sex (p<0.001). Conclusions: AF was a strong predictor of mortality for all causes and cardiovascular mortality in primary care population with increased risk in women.