• Camelia C Scarneciu, Livia Sangeorzan, Mihaela Popescu, Vlad D Scarneciu, and Ioan Scarneciu.
• Camelia C. Scarneciu, MD, PhD. Associate Professor, Fundamental Disciplines, Clinical Prevention Department, Transilvania University of Brasov, Romania.
• Pak J Med Sci. 2018 May 1; 34 (3): 583-589.

ObjectivesTo establish a possible relation of dependency between pulmonary hypertension (PHT) and several factors, with the evaluation of their predictive potential, in Graves' disease.MethodsFor identifying the factors implied in producing PHT and for evaluating its reversibility, we made echocardiography exams, sessions of monitoring the blood pressure during 24 hours and biological test in a group of 42 patients with Graves' disease (group H), comparing them with themselves in a euthyroid status (group E, n=25) and with a control group (group C, n=25). In order to analyse the relation of dependency between pulmonary hypertension (PHT) and the factors identified in the H group, we used both the simple linear regression method (polynomial of degree 1) and the non-linear regression method (polynomial of degree 2, 3) for establishing one model of functional dependency. We used the values of the coefficients of correlation r (degree of dependency) and of determination R2 (the type of dependency). The statistical test (F-test, AIC criterion, test t) was applied by choosing the most appropriate model of determination, with a higher predictive potential.ResultsWe identified PHT at 47.6% of the patients with Graves' disease. Once the euthyroidism status is obtained, PHT is normalized. While inducing PHT, we identified a strong relationship of dependency on several possible new factors such as: pre-treatment period, age, level of the thyroid stimulating hormone receptor antibody and values of systolic blood pressure, besides the already known ones (high level of thyroids hormones, cardiac output, pulmonary vascular resistance).ConclusionsThe non-linear model best explains the relation of determination between pulmonary pressure and those factors having a better predictive potential (from 51% to 90%), compared with the linear model, the only exception being the age factor and the systolic blood pressure, where both models seems to be appropriate.

3 keywords...

hide…