Dysregulation of Plasma Growth Factors and Chemokines in Cocaine Use Disorder: Implications for Dual Diagnosis with Schizophrenia and Antisocial Personality Disorder in an Exploratory Study

Abstract

Introduction: Dual diagnosis in individuals with cocaine use disorders (CUDs) presents a mental health challenge marked by an increased susceptibility to disabling morbidities and premature mortality. Despite extensive research on depression and anxiety, other prevalent comorbidities, such as psychotic and personality disorders, have received less attention. This study explores inflammation-related mediators as potential biomarkers for CUD and dual diagnosis with schizophrenia (SCZ) or antisocial personality disorder (APD).

Methods: This exploratory study included 95 participants, comprising 40 healthy subjects and 55 abstinent patients with CUD. Lifetime CUD was diagnosed either as single diagnosis (CUD group, N = 25) or as a dual diagnosis (DD group. N = 30) with SCZ (CUD+SCZ subgroup) or APD (CUD+APD subgroup). Participants were clinically assessed, and the plasma concentrations of growth factors (i.e., G-CSF, BDNF, and VEGF-A) and chemokines (i.e., CCL11/eotaxin-1, CCL2/MCP-1, and CXCL12/SDF-1) were determined and log(10)-transformed for analysis.

Results: Growth factors and chemokines were dysregulated by CUD and psychiatric diagnoses. Specifically, patients in the CUD group exhibited significantly lower concentrations of G-CSF and CCL11/eotaxin-1 than the control group. In contrast, the DD group showed significantly higher concentrations of all analytes than both the CUD and control groups. Additionally, no differences in these analytes were observed between the CUD+SCZ and CUD+APD subgroups within the DD group. Regarding cocaine-related variables, significant associations were identified in the CUD group: an inverse correlation between the age at first cocaine use and the concentrations of BDNF and CCL2/MCP-1; and a positive correlation between the duration of the cocaine abstinence and the concentrations of BDNF and CCL11/eotaxin-1. Lastly, a logistic regression model incorporating all these analytes demonstrated high discriminatory power in distinguishing patients with CUD alone from those with dual diagnosis.

Conclusions: Individuals with dual diagnosis of CUD exhibit elevated concentrations of growth factors and chemokines, distinguishing them from those with CUD alone. It is unclear whether the differences in these inflammatory mediators are specific to the presence of SCZ and APD. The study highlights potential biomarkers and associations, providing valuable insights into the intricate interplay of CUD and psychiatric disorders to enhance clinical diagnosis and therapeutics.

Keywords: Antisocial personality disorder; BDNF; Chemokines; Cocaine; Dual diagnosis; Eotaxin-1; Growth factors; Schizophrenia.

Fig. 1.

Plasma concentrations of growth factors based on type of CUD diagnosis. G-CSF (a), BDNF (b), and VEGF-A (c). Bars are estimated marginal means and 95% CI of log(10)-transformed concentrations. Data were analyzed using one-way ANCOVA and controlling for age and sex. (***) p < 0.001 and (**) p < 0.010 denote significant main effects of the group factor. (⁺⁺⁺) p < 0.001 and (⁺) p < 0.05 denote significant differences compared with the control group. (^###) p < 0.001 and (^#) p < 0.05 denote significant differences compared with the CUD group.

Fig. 2.

Plasma concentrations of chemokines based on type of CUD diagnosis. CCL11/eotaxin-1 (a), CCL2/MCP-1 (b), and CXCL12/SDF-1 (c). Bars are estimated marginal means and 95% CI of log(10)-transformed concentrations. Data were analyzed using one-way ANCOVA and controlling for age and sex. (***) p < 0.001 denotes significant main effects of the group factor. (⁺⁺⁺) p < 0.001 denotes significant differences compared with the control group. (^###) p < 0.001 denotes significant differences compared with the CUD group.

Fig. 3.

Correlation between plasma concentrations of analytes and cocaine-related variables in the CUD group. a Age at first cocaine use and BDNF concentration. b Age at first cocaine use and CCL2/MCP-1 concentration. c Duration of cocaine abstinence and BDNF concentration. d Duration of cocaine abstinence and CCL11/eotaxin-1. Pearson correlation was performed using log₁₀-transformed concentrations. Dots are individual values. (r) Pearson’s correlation coefficient.

Fig. 4.

Plasma concentrations of growth factors based on comorbid psychiatric diagnosis. G-CSF (a), BDNF (b), and VEGF-A (c). Bars are estimated marginal means and 95% CI of log(10)-transformed concentrations. Data were analyzed using one-way ANCOVA and controlling for age and sex. (***) p < 0.001 and (*) p < 0.05 denote significant main effects of the (sub)group factor. (⁺⁺⁺) p < 0.001 and (⁺) p < 0.05 denote significant differences compared with the CUD group.

Fig. 5.

Plasma concentrations of chemokines based on comorbid psychiatric diagnosis. CCL11/eotaxin-1 (a), CCL2/MCP-1 (b), and CXCL12/SDF-1 (c). Bars are estimated marginal means and 95% CI of log(10)-transformed concentrations. Data were analyzed using one-way ANCOVA and controlling for age and sex. (***) p < 0.001 denotes significant main effects of the (sub)group factor. (⁺⁺⁺) p < 0.001 denotes significant differences compared with the CUD group.

Fig. 6.

Receiver operating characteristic analysis for a logistic regression model to distinguish patients with CUD alone from those with dual diagnosis (CUD and SCZ or APD). a Scatter plot of the predictive probabilities. b Receiver operating characteristic curve generated with probability values from a binary logistic model including log(10)-transformed concentrations of G-CSF, BDNF, VEGF-A, CCL11/eotaxin-1, CCL2/MCP-1, and CXCL12/SDF-1. Lines on the scatter plot are median and IQR. (***) denotes significant differences using the Mann-Whitney U test.