Trial Information

Methylphenidate for Depressed Cancer Patients in Hospice

Background: Major depressive disorder can be diagnosed in between 5% and 26% of terminally
ill patients. This disorder causes suffering, and is associated with suicidality, increased
pain, and increased caregiver burden and caregiver depression. Treatment of depression in
cancer patients in hospice and palliative care is complicated by shortened life expectancy.
Currently-approved antidepressants take several weeks to be effective. Methylphenidate has
been reported in case series and very small randomized trials in patients without cancer as
a rapidly effective treatment for depression in medically ill patients. There are no
randomized controlled trials to test this agent in terminally ill cancer patients.

Objectives: (1) To determine the effectiveness and safety of methylphenidate for depression
treatment in cancer patients receiving hospice and palliative care, (2) to explore whether
successful treatment of depression is associated with improved quality of life, and (3) to
explore whether effective treatment of depression influences caregiver depression and
caregiver burden.

Methods: We will conduct an 18-day randomized, double-blind, fixed-dose (10 mg bid),
placebo-controlled clinical trial of methylphenidate for depression in eligible veteran and
non-veteran cancer patients with advanced cancer in the following settings: inpatient and
outpatient hospice, inpatient and outpatient palliative care, and inpatient and outpatient
cancer clinics. We will determine whether improvement in depression is mediated by decreased
pain and document the safety and tolerability of methylphenidate in these patients. We will
explore whether improvement in depression results in improved quality of life for these
patients, and decreases caregiver depression and burden.

Eligible patients who answer yes to the question "are you sad or depressed" will be invited
to participate. They will complete measures of depression [Structured Clinical Interview for
Diagnosis (SCID), Montgomery-Asberg Depression Rating Scale (MADRS) as primary outcome,
Hospital Anxiety and Depression Scale (HADS) as secondary outcome], quality of life, pain,
and cognition at baseline. MADRS scores must be greater than 19 and SCID positive for
depression at study entry. Subjects will be randomized to either methylphenidate plus an
SSRI, or placebo plus an SSRI. Subjects may continue any previously prescribed SSRI, or will
be prescribed citalopram if untreated. Participants will be evaluated with the same measures
as baseline on days 3, 6, 12 and 18 of the study. In an open label portion of the study,
methylphenidate-treated patients whose depression has improved will be followed up to 2
months. Cox proportional hazard analysis will be used to analyze the primary outcome. An
estimated 104 subjects will be entered over five years. Caregivers will complete measures of
depression and caregiver burden at days 0 and 18.

Findings: Forty-six subjects were enrolled. Because enrollment was lower than anticipated,
we added all cancer clinics at OHSU to increase subject enrollment. Seventy-eight percent of
subjects were men, and their mean age was 64 years.

The mean time to remission of depression was 10.3 days (SE = 1.77) in the methylphenidate
group and 8.1 (SE = 1.31) for the placebo group (p = 0.38, log rank test). The response to
placebo was high, suggesting that even with a larger number of patients our original
analytic approach would not have been able to show a difference. For example, by day six of
the study 69% of placebo patients and 54% of the methylphenidate subjects no longer met
depression criteria. However, after first remission, 5 placebo patients relapsed, where as
only 1 methylphenidate patient relapsed to depression. By day 18, 84.6% of methylphenidate
patients were in remission, compared to 60% of placebo patients (p = NS). On the HADS the
mean score for each group was 10.4 on screening, but had declined to 6.8 on day 18 in the
methylphenidate group and 8.1 on day 18 in the placebo group (p = NS).

Because of the correlated nature of the data, we tested for linear trend in the MADRAS
scores over time using Generalized Estimation Equation modeling. The overall test for linear
trend in both the placebo and methylphenidate groups revealed a significant decreasing
linear trend (p<0.0001). When restricting the analysis to the placebo group only, the linear
trend test revealed a significant decreasing trend (estimated mean = -1.05; p = 0.0002).
When restricting the analysis to the methylphenidate group only, the linear trend test also
revealed a significant decreasing trend (estimated mean = -1.57; p = 0.0007). The mean
decrease was slightly higher in the methylphenidate group.

Status: Project is complete.

Impact: Alternative study design may be needed to determine the effectiveness of
psychostimulants for depression in advanced cancer.