To construct and validate models of risk factors for depression in patients with hypertensive intracerebral hemorrhage (HICH) after surgery.

The study included the case data of 200 patients with HICH who were admitted to our hospital from July 2017 to March 2020. They were divided into postoperative depression group (n=89) and non-depression group (n=111), a univariate analysis was performed on patients. According to the post stroke depression diagnostic criteria, the postoperative depression group was divided into two groups: mild and moderate-severe. Logistic regression analysis was performed on the two groups, a moderate to severe depression risk prediction model was constructed and verified, and the nomogram was constructed with RStudio software.

The incidence of depression in 200 patients with HICH was 44.5%, of which the incidence of mild depression was 59.6%, and the incidence of moderate to severe depression was 40.4%. A univariate analysis of the depression group and the non-depression group showed that, the gender, personality, family income, degree of bleeding, level of paralysis, brain-derived neurotrophic factor (BDNF) level, homocysteine (Hcy) level, nuclear factor erythroid-2 related factor 2 (Nrf2) level of the two groups were statistically significant (P<0.05). Univariate analysis of mild and moderate-severe depression showed that family income, bleeding degree, paralysis level, Hcy level, Nrf2 level, and BDNF level were statistically significant (P<0.05). The inclusion of multivariate Logistic regression analysis showed that the degree of bleeding, the level of paralysis, the level of Hcy, the level of Nrf2, and the level of BDNF were independent risk factors for different degrees of depression after HICH (P<0.05). The sensitivity of the constructed moderate to severe depression risk prediction model was 89.6%, the specificity was 84.6%, and the AUC (95%CI) was 0.794 (0.694-0.831), and the verification results were the same as the constructed results.

The main risk factors for different degrees of depression after HICH are the degree of bleeding, the level of paralysis, the level of Hcy, the level of Nrf2, and the level of BDNF. The risk prediction model constructed by the research can become a risk prediction tool for the occurrence of moderate to severe depression after surgery.