PRACTICAL: A COMPARISON OF TWO STUDIES
N-methyl-D-Asparate (NMDA) for Depression
OBJECTIVES


Understand the design features of randomised clinical trials to evaluate interventions
Understand the main sources of bias in randomised clinical trials
BACKGROUND
Major depression is a serious mental illness that often does not respond to mainstream drug
treatment (antidepressants). In addition, there is usually a delay of 2-6 weeks before mood improves
significantly. In situations like this, when at least two conventional antidepressants have been tried
without success, depression is considered treatment-resistant
SECTION 1: ZARATE TRIAL
Zarate CA Jr, Mathews D, Ibrahim L, Chaves JF, Marquardt C, Ukoh I, Jolkovsky L, Brutsche NE, Smith
MA, Luckenbaugh DA. A randomized trial of a low-trapping nonselective N-methyl-D-aspartate
channel blocker in major depression. Biol Psychiatry. 2013 Aug 15;74(4):257-64.
Ketamine was shown to lead to rapid and sustained antidepressant relief even in treatment-resistant
patient. Please consider the following questions that address the study design and results
Question 1: What type of study was this?
Randomised Clinical Trial
Question 2: What study design did they use?
Cross over
Question 3: What were the inclusion / exclusion criteria? What potential effect would this have on
the generalizability of the results?
Homogenous trial population with strict exclusion criteria. Not very generalizable results.
Question 4: Do you think there was a potential bias through the blinding procedures in place? If so
how?
The study used an inactive placebo (i.e. a saline injection), which is problematic as ketamine
produces strong dissociative side effects. As a consequence, it is likely that participants were able to
infer whether they had been given ketamine or the placebo, which raises concerns regarding
blinding.
Question 5: The researchers completed three different analyses? Which of these was the most
important? What conclusions could you draw if the completers analysis had differed from that of the
intention to treat?
Intention to Treat is the most informative
Completers must have both phases of the cross over study so is subject to attrition bias
Question 6: What was the trial sample size? Were any of the following parameters described in the
study:
JMH 15/07/2014. Dept. of Biostatistics
Power calculation
Clinically meaningful difference
Primary outcome measure
Primary assessment time point ?
Sample size details are not provided
Question 7: What are the potential problems from a small sample size?
Makes generalizability difficult subject to large random error, imprecise estimates.
Increases chances of type I and II errors.
SECTION 2: MURROUGH TRIAL
Murrough JW, Iosifescu DV, Chang LC, Al Jurdi RK, Green CE, Perez AM, Iqbal S, Pillemer S, Foulkes A,
Shah A, Charney DS, Mathew SJ. Antidepressant efficacy of ketamine in treatment-resistant major
depression: a two-site randomized controlled trial. Am J Psychiatry. 2013 Oct 1;170(10):1134-42.
A new study published in the American Journal of Psychiatry has gone some way to addressing some
of the shortcoming in previous N-methyl-D-Asparate (NMDA) for depression research
Question 8: Does this trial address the shortcomings in the Zarate trial?
Shortcomings were: Lack of generalizability, Small sample size, Unsuitable placebo, cross over
design?
Yes
Question 9: What measure did they use for the primary outcome? Why do you think this was chosen
over the primary outcome in the previous study?
HDRS
Psychiatrist measured compared to the user / research rated HAMD in the earlier trial
Question 10: Why was the randomisation allocation at 2:1 rather than 1:1?
The most efficient randomisation is 1:1, power is lost when this ratio deviates
A large effect size was expected; hence the researchers were in a position to use an increased
allocation to the active arm.
Research indicated that the active arm was effective therefore; the study would be more beneficial
to those participants randomised if more were given the active arm.
Ethically more sound.
Might increase recruitment rate, favourable chances of receiving active treatment
Question 11: The clinical assessments were performed by trained, blinded raters. What source of
bias is this designed to avoid?
Systematic error // Assessor bias
Question 12: Interpret the results of this study? Can you spot any limitations to their follow-up
method?
This study only assessed the effects of a single injection of ketamine. While effective in acute
intervention, another main problem with current treatment is how to maintain mood improvement.
From this perspective, it would have been interesting to study the effects of consecutive treatments,
for instance on a weekly basis or treatment with Ketamine followed up with antidepressant
medication over a longer trial period.
JMH 15/07/2014. Dept. of Biostatistics
Thiis point also highlights how little is known about the long-term effects of ketamine treatment.
Also, it should be kept in mind that due to possible dissociative effects, ketamine will be restricted to
non-psychotic forms of depression. While making sense clinically speaking, this is somewhat
unfortunate, as psychotic depression is usually more severe and difficult to treat, making such
patients more likely to be treatment-resistant.
From a methodological perspective, this study raises some doubts concerning the appropriateness of
midazolam as an active control. While 17% of patients reported dissociative symptoms with
ketamine, this was not found for midazolam. It would have been interesting to assess directly (i.e.
via questionnaires) if participants were aware of their experimental condition after the infusion to
rule out these concerns.
JMH 15/07/2014. Dept. of Biostatistics