Depression in glioma

Abstract

BACKGROUND Few high-quality observational studies have been conducted to examine clinically relevant features of emotional distress and Major Depressive Disorder (MDD) in adults with primary cerebral glioma. Our knowledge of these important complications of glioma is currently poor. AIMS This thesis aims to answer a series of relevant clinical questions. I have studied: [1] the frequency, independent clinical associations and course of general emotional distress measured using the NCCN Distress Thermometer (DT); [2] the utility of three depression screening tools for identifying MDD; [3] the frequency, independent clinical associations and course of MDD in glioma; [4] current patterns of practice, and the apparent tolerability of antidepressant treatment of depression in glioma; and [5] barriers to the effective management of MDD in glioma. METHODS I conducted a prospective, twin-centre, observational cohort study. Adults with a new histological diagnosis of primary supratentorial glioma were enrolled and interviewed three times: shortly after starting radiotherapy (T1), three months later (T2) and six months later (T3). At each time point participants completed the DT, the Hospital Anxiety and Depression Scale (Depression subscale, HAD-D), the Patient Health Questionnaire-9 (PHQ-9) and the Structured Clinical Interview for DSMIV MDD (SCID). Barriers to depression management were studied using questionnaires completed by the patient and their named GP. RESULTS During a two-year recruitment period, 223 patients were eligible and 155 provided useable data (57.4% male, mean age = 54.2 years, 85.8% high-grade glioma, 78.1% radical radiotherapy, 55.5% chemotherapy). [1] High distress (DT score ≥ 4/10) was consistently a frequent complication, occurring in between 36.4% ± 7.6% of patients at T1 to 33.7% ± 10.2% at T3. In a logistic regression analysis, high distress at T1 was independently associated with MDD, functional impairment and younger age (χ2 for model = 39.882, p < 0.001, R Square = 0.312). Patients who reported high distress at T1 (median DT score = 8; IQR 7 - 9) remained highly distressed on follow-up (T2 median score = 8, IQR 6 - 8; T3 median score = 7, IQR 5 - 8). [2] As screening tools, the HAD-D and PHQ-9 showed good internal consistency (α = 0.769 - 0.862 at any time point). The HAD-D displayed the best operating characteristics on ROC curve analysis. At a threshold of 7+, sensitivity = 0.933, specificity = 0.907 and Positive Predictive Value (PPV) = 0.56. A threshold of 8+ displayed similar PPV, however. [3] The cross-sectional prevalence of MDD was 13.5% ± 5.4% at T1, 14.8% ± 6.7% at T2 and 6.8% ± 5.8% at T3. Inter-rater diagnostic agreement was good (κ = 0.81, 95% CI 0.60 – 1.00). MDD was independently associated with a past history of depression (OR = 3.8, 95%CI 1.5 - 9.8), and with current functional impairment (OR = 3.6, 95%CI 1.4 - 9.4). MDD persisted for at least three months in 9/17 patients who could be followed up. [4] The frequency of antidepressant prescription was 8.4% ± 4.4% at T1, 7.4% ± 4.9% at T2 and 12.6% ± 6.9% at T3. Citalopram was the most frequent antidepressant choice. Antidepressant tolerability appeared to be good among patients who could be followed up. [5] Barriers to the management of depression included 78.4% of GPs regarding major depression as a normal reaction to having glioma, and 39.2% expressing a belief that major depression did not always require treatment. In addition, most patients expressed a degree of resistance to any kind of future depression treatment. DISCUSSION This is the largest cohort study of depression in consecutively presenting adults with glioma, and the first to utilise criterion standard structured interview diagnoses in a longitudinal design. There is a degree of theoretical uncertainty about the nosological validity of MDD in glioma, although the clinical relevance of this uncertainty can be debated. Methodological limitations to the presented study include an absence of alternative potential psychiatric diagnoses to MDD, the likelihood of selection bias in recruitment, and considerable attrition. Due to these and other limitations, findings from this study are tentative and should ideally be replicated. Clinicians should have a high index of suspicion for identifying low mood in glioma patients, particularly those with functional impairment or previous depressive episodes. The HAD-D (suggested threshold 8+) can reasonably be used to screen for depression, if desired. Caution is required when prescribing antidepressants. Clinicians should be educated about the frequency and consequences of MDD in glioma. Researchers interested in psychological neuro-oncology could convene a meeting to guide future projects, particularly since multi-centre studies may be necessary to recruit sufficient sample sizes in future.