Left ventricular hypertrophy in the automatic electrocardiogram report and mortality risk

Abstract Introduction Left ventricular hypertrophy (LVH) is associated with complex structural changes in the myocardium, altering the electrocardiogram (ECG). The ECG is the initial test for patients with suspected heart disease. Electrocardiographic criteria for LVH diagnosis have a low sensitivity compared to the echocardiogram (ECHO). The advent of tele-ECG and the availability of automatic analysis systems have made the large-scale use of electrocardiography possible. However, there are only a few studies on the prevalence and prognosis of LVH automatically detected in the ECG. Objectives Evaluate the association between the Glasgow program score for LVH in the ECG and overall mortality in an electronic cohort of primary care patients in Brazil. Methods Patients from the CODE (Clinical Outcomes in Digital Electrocardiology) cohort, older than 18 years, who underwent digital ECG from 2010 to 2017, were included. The Glasgow University ECG Automated Interpretation program reported ECGs. LVH in the ECG was classified into definite LVH if the score was≥ 6.3, probable LVH if the score was between 5.0 and 6.3, possible LVH if the score was between 4.0 and 5.0, and no LVH if the score was 4.0. To assess the relationship between LVH-Glasgow score and mortality, Cox regression adjusted for age, sex, and comorbidities was used. Results The CODE database included 1,558,415 patients and 1,389,331 patients over 18 years. Isoelectric ECGs with interference, exchange, or lousy positioning of electrodes, which could compromise the analysis, and those with electrocardiographic diagnoses that impair the diagnosis of LVH (pacemaker rhythm, left bundle branch, and ventricular pre-excitation) were excluded. The prevalence of abnormal LVH-Glasgow score (≥4.0) was 18.5%. At a median follow-up of 3.47 years, the general population’s all-cause mortality rate was 2.68%. After multivariate adjustment, the presence of definite LVH had a higher risk of overall mortality (95% CI; HR 1.64 (1.59-1.69)); probable LVH (95% CI; HR 1.18 (1.14-1.23)) and possible LVH (95% CI; HR 1.09 (1.05-1.13)) were also associated with increasing risk of death. Conclusions The LVH-Glasgow score can be a prognostic tool in ECG analysis. In this population, a higher score was associated with a higher risk of overall mortality.