Better—but good enough? The first randomised controlled trial of psychostimulants for depressive symptoms in advanced cancer
- Department of Psychological Medicine, Institute of Psychiatry, Weston Education Centre, 10 Cutcombe Road, Denmark Hill, London SE5 9RJ, UK
- Correspondence to Dr Lauren Rayner, Department of Psychological Medicine, Institute of Psychiatry, Weston Education Centre, 10 Cutcombe Road, Denmark Hill, London SE5 9RJ, UK;
Depression is common in palliative care, and its adverse impact on quality of life, function and physical symptom severity is well documented.1–3 Alleviating low mood mitigates the emotional, social and physical impact of advanced illness—but for patients with a short prognosis the options are limited. Most psychotherapeutic approaches require a series of sessions delivered over several weeks, and antidepressant medication may take 1–6 weeks to have full therapeutic effect. Recently, there has been a resurgence of interest in the use of psychostimulants for treating depression, especially in patients with advanced disease. The potential for psychostimulants to enhance mood, energy and arousal within hours holds particular promise for palliative patients. Though the evidence base is scant and methodologically shaky,4 several authorities have recommended psychostimulants for patients with a short life expectancy,5 ,6 and they are increasingly being used, particularly in the USA.
This issue presents the results of METIORON—a multicentre, double-blind, randomised placebo-controlled trial of the effect of methylphenidate on depressive symptoms in advanced cancer. Centeno and colleagues’ randomised controlled trial (RCT) is, to our knowledge, the first to evaluate the effect of psychostimulants on palliative patients’ mood as the primary outcome. The study investigators compared response with methylphenidate versus placebo over a 28-day period. In common with many RCTs undertaken in this setting, poor accrual and high attrition undermined the power of the study to provide a definitive answer to the question posed. Nevertheless, the observed non-significant trend towards improved mood with methylphenidate versus placebo is an important finding.
Studies assessing the impact of psychostimulants on depression in non-palliative populations have yielded mixed results, with no clear indication of efficacy. A Cochrane review by Candy et al7 in 2008 identified 24 trials of psychostimulants for depression, but only eight placebo-controlled studies had data suitable for meta-analysis. Five of these were published pre-1980, and all of them omitted key information on the method of randomisation and allocation concealment. Meta-analysis of three trials assessing short-term change in depression score (<5 weeks) showed a significant difference favouring psychostimulants (Standardised Mean Difference (SMD) −0.87, 95% CI −1.40 to −0.33; three trials 62 participants), which exceeds criterion for clinical significance suggested by the National Institute for Health and Clinical Excellence (SMD≥0.5).8 However, meta-analysis of three trials comparing rates of clinical response found no difference between intervention and control conditions (OR 1.01, 95% CI 0.48 to 2.09; three trials; 130 participants). Subgroup analysis of patients with concomitant medical illness also showed no statistically significant effect (SMD −0.75, 95% CI −1.39 to −0.11; two trials; 42 participants). In the short term (<5 weeks), there was no difference in the proportion of patients reporting side effects in the intervention and control arms (two trials; 185 participants), but in the medium term (5–12 weeks), side effects were much more frequently reported by patients taking a psychostimulant (OR 7.22, 95% CI 2.21 to 23.57; two trials; 90 participants).
In advanced disease, the evidence is scarcer still. This is a difficult issue to address using RCT methodology, and until this year, the literature consisted largely of uncontrolled studies and case series. Problems recruiting and retaining trial participants often arise in palliative settings, and may be magnified when the intervention under study is considered controversial. It is notable, however, that psychostimulants have been successfully trialled for the treatment of cancer-related fatigue. A recent systematic review (Milton et al 2011) identified five randomised placebo-controlled trials conducted since 2006.9 Though the sample sizes were modest (n=50–152), and several studies showed no effect, meta-analysis yielded a small but statistically significant benefit over placebo (SMD −0.28, 95% CI −0.48 to −0.09; p=0.005; 426 participants). This burgeoning evidence base throws into relief the dearth of data on depression.
A systematic review in 2009 found no controlled studies of methylphenidate for depression in palliative care (there were six assessing its effect on fatigue).10 Three prospective uncontrolled trials11–13 and two retrospective case series did evaluate depression,14 ,15 but small sample sizes (n=15–41), conflicting findings (response rates ranged from 27 to 81%), and poor methodology precluded inferences about efficacy. More recently, Kerr and colleagues recruited 30 hospice patients to an RCT of methylphenidate for fatigue, with depression assessed as a secondary outcome.16 Between baseline and follow-up (day 14) there was a significant reduction in depressive symptoms in the intervention group, but not in the control group. However, the trial only included 30 patients, and between-group analyses were not reported.
The sample size of the METIORON trial is more than twice that of Kerr and colleagues’ study. Sixty-nine patients with advanced cancer were recruited from eight palliative care centres in Spain. All participants had a life expectancy longer than 1 month, and none were undergoing systemic cancer treatment. Patients were required to score at least 2 out of 8 on the Two Question Screening Survey for depression (which asks about low mood and loss of interest over the last month)—but notably, those with a clinical diagnosis of depression were excluded from enrolling. Participants randomised to the control group (n=38) received a placebo, while those randomised to the intervention group (n=31) received methylphenidate—a central nervous system stimulant, structurally related to amphetamine. Doses of both the active drug and placebo began at 15 mg, increasing to 45 mg at subsequent consultations if the patient showed no response. Participants were assessed five times over a 28-day period (0, +7, +14, +21, +28), and clinically relevant response was defined as an improvement of 2 or more points on the Edmonton Symptom Assessment Scale—a visual analogue scale from 0 to 10. An intention-to-treat analysis of efficacy and adverse events was performed including all participants: 14/31 (45%) responded to methylphenidate, and 10/38 (26%) responded to placebo (difference 19%; 95% CI −4% to 39%; p=0.10). The proportion of patients reporting at least one adverse event was the same in the two treatment arms (71%; p=0.99), but the proportion of patients experiencing three or more events was higher in the methylphenidate group (35% vs 16%; p=0.05). No serious side effects were reported. The authors conclude that in the doses used in this study, methylphenidate is a safe drug with relatively few harmful side effects (though the study does not address concerns about the long-term safety of psychostimulants). They posit that methylphenidate may have a clinically useful effect in relieving depressive symptoms in advanced cancer patients, but point out that the findings of the current study are inconclusive.
The authors’ interpretation of the study findings is appropriately cautious. Though larger than previous studies, the sample size is still small, and the effect favouring methylphenidate is non-significant. Prior to recruitment it was estimated that 98 participants (49 in each study arm) would be required in order to detect a difference in response. Yet despite extending the enrolment period from 12 to 28 months, only 69 patients were recruited into the trial, and only 19 completed the final assessment on day 28. The authors speculate about the reasons for poor recruitment, but regrettably did not collect data on non-participation (eg, ineligibility, possible ‘gate-keeping’ by clinical staff reluctant to expose their patients to research, and refusal). The criterion used to define response to treatment (improvement of two or more points on the ESAS) was unusual and makes comparison with trials using the conventional criterion (≥50% decrease in depressive symptom score) difficult. Another limitation is the exclusion of patients with depression who might be eligible for an antidepressant, presumably an ethical requirement owing to the stronger evidence base for these drugs. However, we suggest that if the benefits of psychostimulants were their rapid onset of action for patients whose life expectancy is measured in weeks rather than months, there is a sound rationale for a short-term trial of psychostimulants in patients with major depression. Unfortunately, the present trial can comment only on the effect of methylphenidate on depressive symptoms in advanced cancer; which may be different from its effect on clinically diagnosed depressive disorder.
Although imperfect, Centeno and colleagues’ study is a welcome addition to the literature, not least as an illustration of the difficulties of conducting trials in this context. The data generated, even if imperfect, are considerably stronger than the alternative method of reporting open label case series. The study suggests investment in a larger RCT of methylphenidate for the rapid improvement in mood in patients with depression is warranted. Until such a study is conducted, psychostimulants cannot be recommended for the treatment of depression in palliative care.
Contributors LR planned, drafted and revised the manuscript. MH conceived, and critically revised the manuscript. LR is guarantor.
Competing interests None.
Provenance and peer review Commissioned; internally peer reviewed.
- Received 24 September 2012.
- Accepted 9 October 2012.
- Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://group.bmj.com/group/rights-licensing/permissions