The practical value of a Logistic regression model of serum indexes in differentiating between geriatric depression and depressive state

Abstract

Objective To investigate the practical value of a logistic regression model of serum indexes in distinguishing between geriatric depression and geriatric depressive state. Methods A total of 160 patients were recruited from the outpatient department from January 2013 to January 2016, and were divided into a depression group(n=80)and a depressive state group(n=80), with retrospective diagnoses based on the Chinese Classification of Mental Disorders-Third-Edition(CCMD-3). Serum samples were collected and enzyme linked immunosorbent assays(ELISA)were used to determine the concentrations of brain-derived neurotrophic factor(BDNF), glial cell line-derived neurotrophic factor(GDNF), fibroblast growth factor 2(FGF-2), vascular endothelial growth factor(VEGF), interleukin-1β(IL-1β), interleukin-6(IL-6), interleukin-10(IL-10), tumor necrosis factor-α(TNF-α), cortisol(CORT), and platelet-derived growth factor(PDGF). Binary logistic regression analysis was used to establish the regression model, and the ROC curve was drawn to explore its value of differentiating geriatric depression from depressive state. Results The levels of serum BDNF, GDNF and VEGF in the depression group were lower than those in the depressive state group(BDNF: 208.7±41.4 vs.262.9±84.6 ng/L, GDNF: 92.3±18.6 vs. 101.4±30.9 ng/L, VEGF: 223.1±98.2 vs. 257.8±77.2ng/L), while the levels of IL-1β and CORT in the depression group were higher than in the depressive group(IL-1β: 27.0±4.9 vs.19.6±5.7 μg/L, CORT: 96.3±16.7 vs.83.0±17.3 nng/L). In multivariate logistic regression analysis, BDNF(OR=0.987, P = 0.001), IL-1β(OR=1.29, P = 0.000)and CORT(OR=1.065, P = 0.000)were selected to build the regression model.The regression equation was P=1/[1+ e -(- 8.546 - 0.013(BDNF)+ 0.258(IL-1β)+ 0.063(CORT))]and the area under the ROC curve was 0.966.Compared with retrospective diagnoses made two weeks later, the correct diagnosis rate of the logistic model was 90.47%. Conclusions The Logistic regression model of serum indexes can further differentiate between geriatric depression and depressive state which also offers additional benefits for the diagnosis, differential diagnosis, and treatment of depression. Key words: Hamilton's depression scale; Logistic model; ROC curve; geriatric depression; depressive state