MDDScore

Abstract

Background: Previously, a biomarker panel was developed for use as an aid to major depressive disorder (MDD) diagnosis; it consisted of 9 biomarkers associated with the neurotrophic, metabolic, inflammatory, and hypothalamicpituitary-adrenal axis pathways. This panel and associated algorithm produced good clinical sensitivity and specificity (92% and 81%, respectively) in differentiating MDD patients from individuals without MDD. To further validate the panel, we performed a prospective study using a larger set of new prospectively acquired MDD patients and a similarly collected population of nondepressed subjects. The addition of gender and body mass index (BMI) effects to the algorithm was also evaluated.Method: Blood samples were obtained from MDD patients (n = 68) clinically evaluated at multiple sites in 2011 and 2012 using standard psychiatric assessment tools and structured clinical interviews according to DSM-IV criteria. Blood samples (n = 86) from nondepressed subjects were obtained as controls. MDD and nondepressed samples were randomized into independent training (n = 102) and validation sets (n = 52). Analytes in sera were quantified by immunoassay. Results:Training set biomarker data were used to develop a logistic regression model that included gender and BMI in a manner that allowed for their interaction with the biochemical analytes. For the training set, the sensitivity and specificity of the test (with 95% CI) were 93% (0.80-0.98) and 95% (0.85-0.99), respectively. This method (designated the MDDScore) was then applied to the independent validation set and had a sensitivity and specificity of 96% (0.77-0.98) and 86% (0.66-0.95), respectively. The overall accuracy for the training set was 94%; the validation set accuracy was 91%. Conclusion:Examination of a randomized independent set of samples confirms the ability of the previously established biomarker panel to identify persons with MDD; the accuracy was over 90%. The improved model that adds gender and BMI to the previously established panel of 9 biomarkers is robust and simple; it provides the most rigorously tested, objective diagnostic test for MDD to date.

Identification of an Immune-Neuroendocrine Biomarker Panel for Detection of Depression: A Joint Effects Statistical Approach

“…This resulted in a poor performance (AUC = 0.55) for detecting NESDA depression patients and a moderate performance for detecting first-/recent-onset unmedicated patients from specialised mental healthcare (AUC = 0.75; stage 1 cohorts). Recently, Bilello et al [28] published a further study reporting on refinement of the MDDScore algorithm using samples from 68 depression patients and 86 controls with a test sensitivity and specificity >90%. Unfortunately, we were not able to reproduce these findings.…”

“…Recently, Papakostas et al [25] and subsequently Bilello et al [28] reported on development and refinement of a 9-analyte panel (MDDScore test) to discriminate severe recurrent depression patients from controls with a test specificity and sensitivity >90%. However, it has not been tested on naturalistic patient and control populations recruited from the general population and primary healthcare settings, where the need is greatest.…”

Is there Progress? An Overview of Selecting Biomarker Candidates for Major Depressive Disorder

“…In addition to achieving high sensitivity and specificity in their pilot study, Papakostas et al also produced a similar performance in their replication study. This group further refined their model algorithm by factoring in gender, BMI, and normalized cortisol levels ( 219 ). Another group used CSF concentrations of multiple analytes including inflammatory biomarkers (IL-6), serotonin metabolites (5-hydroxyindoleacetic acid), dopamine metabolites (homovanillic acid), and HPA axis biomarkers (hypocretin) to detect severe suicidal behavior and increased risk of completing suicidal attempts in MDD patients ( 220 ).…”

Combined serum levels of multiple proteins in tPA-BDNF pathway may aid the diagnosis of five mental disorders

“…However, late studies showed that a single protein has a little diagnostic value among diverse populations because the results are not always consistent. Recently, studies found that combined different serum proteins would provide a much higher diagnostic value than a single protein 17 – 20 , which indicated that this kind of combination could potentially become a diagnostic tool.…”