Multicentre, double-blind, randomised placebo-controlled clinical trial on the efficacy of methylphenidate on depressive symptoms in advanced cancer patients
- Carlos Centeno1,
- Alvaro Sanz2,
- Miguel Angel Cuervo3,
- Daniel Ramos4,
- Silvia Hernansanz5,
- Jesús Gonzalez6,
- Maria Jose Almaraz7,
- Marcos Lama8,
- Francisco Vara9,
- María Nabal10 and
- Antonio Pascual11
- 1Palliative Care Team, Clinica Universidad de Navarra, Pamplona, Spain
- 2Medical Oncology, Hospital Universitario del Rio Hortega, Valladolid, Spain
- 3Palliative Care Team, Hospital Infanta Cristina, Badajoz, Spain
- 4Palliative Care Team, Hospital Provincial de Zamora, Zamora, Spain
- 5Palliative Care Team, Hospital Clinico Universitario de Valladolid, Valladolid, Spain
- 6Palliative Care Team, Institut Catala Oncologia, Hospitalet, Spain
- 7Palliative Care Unit, Hospital Txagorritxu, Vitoria, Alava, Spain
- 8Palliative Care Unit, Hospital San Juan de Dios, Pamplona, Navarra, Spain
- 9Palliative Medicine Regional Centre, Hospital Universitario de Salamanca, Salamanca, Spain
- 10Palliative Care Unit, Hospital Universitari Arnau de Vilanova, Lleida, Spain
- 11Palliative Care Unit, Hospital Santa Creu i Sant Pau, Barcelona, Spain
- Correspondence to Alvaro Sanz, Hospital Universitario del Rio Hortega, Medical Oncology, Valladolid E-47012, Spain;
Contributors CC: He made substantial contributions to conception and design, acquisition of data, and analysis and interpretation of data; he had an essential role in drafting the article; and in final approval of the version to be published. AS, MAC, DR, SH, JG, MJA, ML, FV, MN and AP: They made a substantial contribution to conception and design of the study, acquisition of data and analysis and interpretation of data; they helped in revising the article critically for important intellectual content; and collaborated in the approval of the final version.
Introduction Methylphenidate is a psychostimulant that has been used to relieve depressive symptoms in advanced cancer patients. No studies compare its efficacy against placebo in this group of patients.
Objective To explore the efficacy of methylphenidate compared with placebo in the relief of depressive symptoms in advanced cancer patients.
Material and methods A multicentre, double-blind, randomised placebo-controlled clinical trial was undertaken comparing the efficacy of methylphenidate and placebo in depressive symptoms. Advanced stage cancer patients were eligible if they scored at least two points on the Two Question Screening Survey for depression. A reduction of at least two points on the Edmonton Symptom Assessment Scale for depression (0–10) was considered as a response.
Results Sixty-nine patients were included (methylphenidate: n=31, placebo: n=38); median daily dose of methylphenidate was 25 mg. Fifty-eight patients (84%) who completed the first week of treatment were considered suitable for evaluation. In the intention to treat analysis, there were 14/31 (45%) responses with methylphenidate and 10/38 (26%) responses with placebo (difference: 19%; 95% CI: 4% to 39%; p=0.10). With the Hospital Anxiety and Depression Scale, 11/19 (58%) patients with methylphenidate and 10/24 (42%) with placebo improved from a score compatible with depression in the first 7 days (difference 16%; 95% CI 13% to 42%; p=0.29). The proportion of patients indicating adverse effects was similar for both cohorts (p=0.99).
Conclusion Compared with the placebo, methylphenidate demonstrated a positive trend in the incidence of response for depressive symptoms in advanced cancer patients.
Approximately 25% of advanced cancer patients suffer from depression.1 However, the proportion of patients with depressive symptoms who do not satisfy the specific criteria of depression seems to be higher.2 These cases are often associated with concepts such as adaptive disorder, dysthymia or emotional distress. In such situations, the therapeutic approach may include emotional support as counselling, or the use of psychoactive drugs as antidepressants or anxiolytics.
To avoid delay in both the identification of, and treatment for, depression in patients whose life-expectancy is limited, there is a trend towards a simple, brief approach based on diagnostic estimation and supported by elementary screening techniques that are reinforced, wherever possible, by more complex tests.3 Within this context, the Two Question Screening Survey (TQSS) for depression focuses on the state of mind and anhedonia and asks if the patient has been feeling low, depressed or despondent, or has felt a lack of interest or satisfaction over the previous month. The TQSS is a valid tool in the screening for depression in cancer patients,4 with high sensitivity and positive predictive value but with a somewhat limited specificity.5 ,6
Once depressive symptoms are presented in a terminally ill cancer patient, palliation of these symptoms becomes an urgent objective, and this will usually involve the choice of rapidly acting treatments.7 Because of delay in the effect of classic antidepressants and because mild depressive symptoms cannot be diagnosed as classic depression, they are not considered as the best chance for these depressive symptoms, above all in patients with a very limited life expectancy. As a result, psychostimulant drugs such as methylphenidate have been considered as an appropriate therapeutic option. These psychostimulants are used to treat depressive symptoms, asthenia and opioid-induced somnolence.8 As opposed to classic antidepressants, psychostimulants start to take effect very quickly, often within hours.9 The percentage of advanced cancer patients with depressive symptoms showing response to methylphenidate ranges from 50% to 75%, with a low incidence of side effects (<20% of this group of patients).10 ,11 ,12 The most frequently reported side effects are anxiety, insomnia and trembling, although withdrawal of treatment due to these side effects is reported in only one out of every eight or ten cancer patients. Even when a systematic review does not support the use of psychostimulants in palliative care because of the marginal efficacy and potentially severe side effects,13 this remains still an open question as data from randomised controlled trials are not homogeneous.14 ,15 This study aims to determine the efficacy of methylphenidate compared with placebo in the treatment of depressive symptoms among advanced cancer patients receiving palliative care without diagnosis of major depression.
Patients, material and methods
METIORON is a multicentre, double-blind, randomised placebo-controlled clinical trial designed to evaluate the efficacy of methylphenidate on depressive symptoms in advanced cancer patients. The protocol was approved by the Ethics Committee for Clinical Research of the Hospitals where the study was undertaken. The clinical trial was registered in EudraCT, the EMEA database for clinical trials (code EUDRACT 2004-001295-39).
The inclusion criteria for taking part in the trial were: advanced cancer with no option of radical treatment; estimated life expectancy of at least 1 month; age equal to or over 18 years; and adequate cognitive function demonstrated in the Mini-Mental Status Examination according to reference data for Spanish population, adjusted for level of education and age.16 Exclusion criteria were: patients receiving antidepressants or ongoing systemic cancer treatment; clinical evidence of psychosis or a history of a major depressive episode (because those patients require more intensive approach); suicidal tendencies; addiction or overuse of psychoactive drugs; severe renal or liver failure; severe heart disease (arrhythmia, angina pectoris, non-controlled hypertension); convulsions; simultaneous use of medications for which methylphenidate is contraindicated; glaucoma; hyperthyroidism; hypersensitivity to methylphenidate; or if the patient had taken part in another clinical trial within the previous 30 days.
Patients were included in the study if they scored at least two points on the Spanish version of the TQSS. The Spanish version was adapted from the English one and includes two questions: ‘Durante el último mes, ¿se ha sentido a menudo bajo de ánimo, deprimido o desesperanzado?’ (‘Over the last month, have you often felt down, depressed or despondent?’) and ‘Durante el último mes, ¿ha sentido a menudo poco interés o satisfacción al realizar cosas?’ (‘Over the last month, have you often felt low interest or satisfaction when doing things?’). The possible answers to both questions and their scores were: ‘Not at all’ (0), ‘A little’ (1), ‘To some extent’ (2), ‘Quite a lot’ (3) and ‘A lot’ (4). All patients who agreed to participate signed an informed consent form prior to inclusion in the study.
A multicentre, double-blind, randomised placebo-controlled clinical trial on the efficacy of methylphenidate on depressive symptoms in advanced cancer patients was undertaken. We did a process of central randomisation that determined, before starting the recruitment, the treatment assigned to every one of patients on study. According to sample size and in order to allow competitive recruitment, a random list of 320 patients was generated (by computer). This list was initially divided into 10 different blocks (for the initial 10 researchers), and each block consisted of eight smaller blocks each containing eight patients. When a new centre entered the study a new researcher's cohort was included. The selection period forecast was initially for 1 year.
The monitoring period for each patient in the study was 28 days. During this time, six visits (at home or in hospital) were undertaken at days 0, +2, +7, +14, +21 and +28, respectively. Additional visits were allowed according to the needs of each patient. During these visits, the dose could be adjusted according to researchers' criteria. The minimum follow-up period for a patient to be considered suitable for assessment of response was 8 days (visit on day +7). In all cases, psychological support for patients with depressive symptoms was rigorously maintained.
Medication and dosage choice
The drug used in the study was methylphenidate (Rubifen), for oral use, provided by Laboratorios Rubió (Barcelona, Spain).The study offered five different dosage levels, with total daily doses ranging between 10 mg and 45 mg (table 1). Each dose of methylphenidate was matched by a corresponding dose of placebo. Dose adjustment was inspired in our clinical practice. Patients entering the study (day 0) received the drug at level 0 (15 mg daily). If in the second visit (day +2) toxicity was evident, the dose was reduced to level -1. If the patient showed no response, the dose was increased to the next level (level +1). If the patient displayed improvement, a stable dose was maintained or an increase to level +1 was implemented. If on re-evaluation at day +7 no response was found, the patient left the study and the evaluation was considered as no response. After day +2, it was left to researchers' discretion to set different ranges in dosage increases. In all cases, at least 2 days should pass before considering a new increase in dosage. These changes in dosage, depending in the response or toxicity, could be undertaken either during the weekly visits planned for in the protocol or during other unplanned additional visits requested by the patient.
An evaluation of the treatment regimen was undertaken using data from all the patients who had received at least one dose of the medication. The analysis included a detailed description of all recent adverse events registered by the researcher, including the nature, intensity and relationship with the medication being studied. The analysis was undertaken using the last observation carried forward. Patients who abandon the study were considered as therapeutic failures.
To monitor depressive symptoms, the visual numerical scale (VNS; 0–10) included in the Edmonton Symptom Assessment Scale (ESAS) was used. This depression rating is regularly used as an outcome indicator in palliative patients.17 Other tools such as the Hospital Anxiety and Depression Scale (HADS) are also research tools but are not routinely used in the daily care of patients due to the lengthy completion processes that may prove burdensome both for the patient and healthcare professional alike. The ESAS is a simple and easy to complete tool validated for advanced cancer patients. It includes a series of VNS examining symptoms in cancer patients (pain, asthenia, anorexia, nausea, somnolence, anxiety, depression, difficulty sleeping, general well-being, breathing difficulties and other symptoms). In this study, an improvement of at least two points in the depression scale of ESAS was accepted as response criteria as this change was deemed to be clinically relevant.
The HADS was designed to detect depression and anxiety in patients admitted to non-psychiatric units. It is a survey focused on psychological issues rather than on physical problems. This is a self-administered survey with 14 items (Likert scale; 0–3) included in two subscales, depression and anxiety. In the validated Spanish version, cut-off values are similar for each subscale: 0–7, normal; 8–10, dubious; and 11 or more, clinical problem. HADS as a combined scale has a single cut-off point of 20 with a sensitivity of 0.77, a specificity of 0.89 and a positive predictive value of 0.48.18
Response to treatment was defined as a decrease of two or more points in the depression VNS included in the ESAS. This size effect was determined by consensus among the clinical researchers involved in the study. The size of ≥2/10 in a verbal or linear numerical scale corresponds with other clinical research studies on symptom relief.19 ,20 ,21 We previewed a response rate to placebo of 40%. In order to detect a difference of more than 30% in the proportion of responses with methylphenidate compared with placebo, 49 subjects were required for each arm of the study on the basis of a bilateral α risk of 0.05 and a power of 80%, with a previewed proportion of lost of follow-up of 20%.
Data were presented within the CI of 95% (95% CI). The quantitative data on continual variables were evaluated with the Student's t distribution where normality was the case, or using the Wilcoxon–Mann–Whitney where it was not. For categorical variables, the χ2 test was used. Statistical analysis was undertaken principally using the SPSS V.15.0 Statistics application (SPSS, Chicago, Illinois, USA).
Between November 2004 and March 2007, 69 patients were included in the study. Recruitment to the study finished before the planned number of 98 patients could be reached. Two Home Palliative Care Units and eight Hospital Palliative Care Units recruited patients to the trial; the median number of patients included per unit was four (limits: 1–26). The researchers were usually the clinicians who routinely treated these patients, and they approached the patients for consent. The outcomes were collected by the same physician or any other associate researcher. Patients followed-up at home do not need to make a visit to a clinic for the scheduled follow-up as it was made at home by the team. A majority of patients were treated at home as a majority of hospital-based units coordinate resources to follow both inpatients and outpatients. The characteristics of patients are shown on table 2.22 Both arms are adequately balanced except in a higher proportion of patients with suspected anxiety in the placebo arm (13/38; 34%) than in the methylphenidate arm (3/31; 10%). A total of 58 (84%) patients who completed at least 7 days of treatment were considered as evaluable. The flow diagram for patients throughout the study is shown in figure 1. Overall survival did not differ between arms: median survival was 37 days (IQR: 22–62 days) in the methylphenidate group and 36 days (IQR: 26–90 days) in the placebo group (p=0.16; log rank).
All 69 patients receive at least one dose of the medication and were therefore included on analysis. All patients commenced with a baseline dose of 0 (methylphenidate 10+5 mg daily or identical amounts of the placebo). In no instances was it necessary to reduce the amounts of the dose. Increased doses were required in 21/30 (70%) patients receiving methylphenidate and in 27/36 (75%) receiving placebo. Median steady dose in the majority of cases in both arms of the study remained level at +1 (15+10 mg daily of methylphenidate or identical amounts of the placebo corresponding to a mean dose 24 mg; limits: 15–45 mg). Median duration of treatment was 3 weeks. Only minor problems were recorded in treatment compliance; forgetting to take or only partially taking the dose was recorded in 5/66 (8%) patients.
In the analysis of the last recorded observation after at least 1 week of treatment, response (a reduction of two points in the VNS of depression) was found in 14/26 (54%) patients receiving methylphenidate and 10/32 (31%) patients receiving placebo (difference: 23%; 95% CI: −3% to 44%; p=0.08). In the analysis of the whole group of patients who receive at least one dose of assigned treatment, there were 14/31 (45%) responses to methylphenidate and 10/38 (26%) patient responses to placebo (difference: 19%; 95% CI: −4% to 39%; p=0.10). The trends in the responses in ulterior visits are shown in table 3 and table 4; the most evident difference favouring methylphenidate is seen in the first week of treatment. Depression and anxiety were also evaluated with the HADS (table 5); no relevant differences were found in the changes from the baseline score between both treatments. Patients who changed in the first 7 days from a score compatible with depression to one of normality or merely diagnostic estimation were 11/19 (58%) with methylphenidate and 10/24 (42%) with placebo (difference 16%; 95% CI: −13% to 42%; p=0.29).
Adverse events were registered in 22/31 (71%) patients receiving methylphenidate, and 27/38 (71%) patients receiving the placebo (p=0.99). In all, 11/31 (35%) patients receiving methylphenidate reported three or more adverse events, compared with 6/39 (16%) patients receiving the placebo (p=0.05). The side effects observed in relation to the treatment are recorded in table 6. All of the side effects presented in less than 10% of patients.
This paper communicates the results of a randomised, placebo-controlled clinical trial evaluating the use of methylphenidate in the relief of depressive symptoms in advanced cancer patients. In spite of extending the recruitment time beyond the initial 12-month period, this trial had a very slow accrual and only managed to recruit 69 patients from the 98 initially expected. There is, therefore, a significant limitation in the power of the study in detecting meaningful results. However, in spite of this limitation, there are still a number of important results that can be derived from this trial.
This study is the first multicentre, double-blind, randomised, placebo-controlled clinical trial to explore the efficacy of methylphenidate to palliate depressive symptoms in advanced cancer patients. While multiple reviews and book chapters recommend the use of methylphenidate and other psychostimulants for depression, there have been no controlled studies. This highlights the difficulty in conducting research in this population, particularly with drugs that are not under patent protection and therefore not likely to receive strong industry funding. Only advanced cancer patients ineligible for any reason for treatment with antidepressants were chosen. In order to conform to ethical requirements in palliative care, emotional support was provided to all patients.
This study had a low accrual rate. Unfortunately, patient refusal rates and associated reasons for refusing to take part in the study cannot be presented, as these data were not rigorously collected. Despite the efforts of researchers to adapt their practice to the present standards for randomised clinical trials, we must recognise that, as in this trial, a relevant proportion of studies on symptom control in cancer patients do not adequately report on how many patients eligible for participation were not approached or did not accept to participate or the characteristics of those patients. A number of practical reasons contributing to the poor recruitment rate are suggested: the reputation of methylphenidate as an amphetamine among patients and even among the researchers themselves; the limited experience in clinical trial and collaborative research among most palliative care teams; a lengthy follow-up period of 4 weeks duration with weekly follow-up may prove a barrier to patient recruitment in the mind of many clinical researchers; and, in particular, exclusion criteria that prevent patients receiving antidepressants or active cancer therapy from taking part in studies. The exclusion criteria of having completed cancer treatment were designed with the aim of relating to a more uniform population of terminally ill cancer patients. All of these limitations should be acknowledged in future trials with drugs on symptoms in palliative care patients. The profile of patients included in this study is similar to those normally treated at palliative care units, with a median overall survival of less than 6 weeks. In all, 11 (16%) of the 69 patients did not complete 7 days of treatment (only four because of clinical deterioration). A significant number of patients withdrew from the study during follow-up, with only 19 (28%) from an initial cohort of 69 completing the preplanned 4 weeks. Three out of four patients failed to complete the originally planned period, but this attrition rate is comparable with other similar palliative care trials; these statistics should be carefully considered when evaluating the results of studies on terminal cancer patients.
As the difference between the arms of the study do not achieve statistical significance (p=0.10), the overall results must be interpreted as inconclusive. In our study, we also used HADS as a secondary outcome. The evaluation profile for depression on the HADS provides a more complete vision than the VNS on the ESAS. In this trial, the overall HADS results also display a trend in favour of methylphenidate consistent with the changes in the depression single item of ESAS (table 4).23 Furthermore, there are indirect data suggesting a higher efficiency for methylphenidate as only 4/31 (13%) patients in the methylphenidate arm abandoned the study due to the inefficacy compared with 12/38 (36%) patients receiving the placebo (p=0.06). A review of trials showed that psychostimulants significantly reduced short-term depressive symptoms in comparison with placebo.
The activity reported (remission rate) in the placebo group is striking. Although the percentage of response per protocol is 26%, this figure rises to 47% in the first evaluation, carried out after 2 days of treatment. Depressive symptoms in patients receiving palliative care are unstable.24 In fact, there are prior data suggesting significant activity within the placebo group in palliative care trials, with responses exceeding 50% in the treatment of asthenia.25 This suggests that palliative care patients may show a higher placebo response in comparison with other patient groups. Although such findings need to be confirmed in further studies, they highlight the need for placebo-controlled studies in this patient population.
During the initial evaluation, some patients reported moderate to severe fatigue. In spite of the existence of data suggesting an effect of psychostimulants on cancer-induced fatigue, no significant differences were found in the evaluations of fatigue carried out during the first week of treatment.26 In addition, no significant changes were observed in appetite or in sleeping difficulties (data not shown) for patients taking methylphenidate. This may be considered as surprising since these factors are often cited as potential side effects of the drug. No significant differences were recorded in the proportion of patients who reported side effects on both treatment arms, although the percentage of patients reporting three or more different side effects was indeed higher among those receiving methylphenidate than in those receiving placebo: 35% versus 16% (difference: 19%; 95% CI: −1% to 39%; p=0.05). In no cases were serious side effects related to methylphenidate recorded. It may therefore be suggested that methylphenidate, in the doses used in this study, is a safe drug with relatively few harmful side effects.
The results in this study also show a certain loss of effectiveness and a reduction in the differences between methylphenidate and the placebo as the trial continues (table 3). This trend towards the convergence in results, which can be interpreted as a manifestation of regression towards the mean, has also been found in other palliative care trials.27 ,28 As any analysis of the impact of immediate effect drugs (such as psychostimulants for depressive symptoms) is subject to changes with longer follow-up, we understood that the ideal time for assessment of depression in this population would be on the first 2 weeks. Longer follow-up implies a in a decrease in initial differences.
To conclude, in this study methylphenidate was well tolerated, a striking placebo effect was evident and a high rate of attrition occurred. Compared with placebo, methylphenidate shows a statistically non-significant positive trend in relieving depressive symptoms in advanced cancer patients. We consider these findings to be inconclusive, and indicate that there may remain clinically useful benefits of methylphenidate, but as these benefits will require a large, carefully conducted trial to be identified, the present study may be interpreted as a pilot study for such a trial.
Funding This study was funded by an unrestricted educational grant from Laboratorios Rubió SA.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
- Accepted 24 April 2012.
- Published Online First 31 May 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