Predictors of neurocognitive impairment at 2 years after a first-episode major depressive disorder

Abstract Background Neurocognitive impairment is a contributor to major depressive disorder (MDD). However, MDD patients show great variability in the level and course of deficits. The present longitudinal study was to identify predictors of neurocognitive impairment in first-episode MDD patients. Methods Neurocognitive performance was analyzed in a cohort of 100 patients at 2 years after a first-episode MDD. Subgroups, deficit type vs. non-deficit type, were compared on baseline clinical, neuropsychological, premorbid and sociodemographic characteristics. The analysis was performed using the multivariate logistic regression to obtain a model for neurocognitive impairment determination. The predicted probabilities of multivariate logistic regression were analysed using Receiver Operating Characteristic (ROC) curve. Results Fifty-two percent of MDD participants presented general neurocognitive impairment. The regression analyses demonstrated that clinical and sociodemographic characteristics were not predictive variables. A model composed of processing speed, executive function, and attention, dexterity correctly classified 85.8% of the MDD patients with deficit type. ROC curve indicated that the changes of these three cognitions could identify MDD with deficit type from MDD with non-deficit type. In addition, ROC curve also indicated that processing speed and executive function could identify MDD from CN subjects. Finally, processing speed performance was negatively correlated with Hamilton Depression Scale scores in both MDD with deficit and non-deficit type. Conclusion The present study provides novel insights on frequency and neurocognitive profile of subtypes of patients showing impairment. Our results suggest processing speed impairment is a trait dimension of the disorder related to specific cognitive dysfunctions and the severity of depression.