Significant outcomes
∙ This exploratory meta-analysis showed a statistically significant effect in favour of deep brain stimulation in treatment-resistant major depressive disorder.
∙ In terms of patients-needed-to-treat, the results show that one out of three patients with long-term treatment-resistant depression can be expected to benefit from deep brain electrical stimulation.
Limitation
∙ Estimates of sham responding to deep brain stimulation are derived by expert opinion rather than by properly controlled clinical trials.
Introduction
Direct electrical stimulation in the brain continues to attract much attention for relief of treatment-resistant depression (Reference Berlim, Mcgirr, Van Den Eynde, Fleck and Giacobbe1,Reference Schlaepfer, Bewernick, Kayser, Hurlemann and Coenen2). There is, however, still great uncertainty concerning efficacy, due in part to lack of information on sham responding to the intensive selection and surgical procedures (Reference Blumberger, Mulsant and Daskalakis3). In the absence of such information, we cannot establish the likelihood of success of the intervention. There are, however, ways of estimating the likelihood of sham responding in certain situations (Reference Merkl, Schneider and Schonecker4). Here, expert opinion is applied to gain an understanding of the potential role of sham responding in the efficacy of deep brain stimulation in patients suffering from long-term treatment-resistant depression.
Materials and methods
Information on the likelihood of sham responding in patients with long-term treatment-resistant depression was solicited from three eminent colleagues (Helen Mayberg, Donald A. Malone Jr., and Thomas E. Schlaepfer) with extended experience using deep brain stimulation for psychiatric disorders (Reference Mayberg, Lozano and Voon5–Reference Schlaepfer, Cohen and Frick7). The information that they supplied along with the present authors’ understanding of treatment-resistant depression (Reference Smith, Stork and Wegener8) placed the likelihood of sham responding in the range 10–25%. Using random number software (http://www.random.org/integers/), nine values were generated and were applied consecutively to the data provided by a recent review on deep brain stimulation in long-term treatment-resistant depression (Table 1) (Reference Schlaepfer, Bewernick, Kayser, Hurlemann and Coenen2).
Summary of data used for an exploratory meta-analysis on benefit of deep brain electrical stimulation in patients with long-term treatment-resistant depression
The information in the first three columns is summarised from a very recent review (Reference Schlaepfer, Bewernick, Kayser, Hurlemann and Coenen2). The estimates of sham response were obtained by expert opinion plus random number software (http://www.random.org/integers/) (see ‘Materials and methods’ section).
Results
Effect sizes were too heterogeneous to uphold a fixed effects model (Q=36.2, df=7, p<0.01) (Reference Lipsey and Wilson9). Therefore, a random effects model was used, and it provided an estimate of effect size of 1.71 (95% CI: 1.47–1.96; Z=13.6, df=7, p<0.01) for deep brain stimulation versus sham treatment in patients with long-term treatment-resistant depression. In the present context, an effect size of 1.71 means that deep brain stimulation was 71% more likely than sham treatment to be of benefit for patients with long-term treatment-resistant depression.
Discussion
The present exploratory finding can be expressed in terms of patients-needed-to-treat (Reference Connolly and Thase10). Assuming a sham response rate of 20% to the intensive selection and surgical procedures required by deep brain stimulation, then we can expect 2 out of 10 patients to report some benefit even in the absence of electrical stimulation. The estimate of effect size derived by the present exploratory meta-analysis indicates that, in addition to sham responding, another 1.4 patients (i.e. sham response×0.71) can be expected to benefit from electrical stimulation. Thus, for every 2.9 patients (10/3.4) with long-term treatment-resistant depression receiving deep brain electrical stimulation, one can be expected to benefit. Whether this level for patients-needed-to-treat is viewed as large or small is, of course, a matter of opinion.
A shortcoming of the present analysis relates to the current lack of empirical evidence on the level of sham responding to deep brain non-stimulation in patients with long-term treatment-resistant depression. Information on that topic can be expected to come eventually from properly controlled, large-scale clinical trials that either disprove or confirm the present findings. One reviewer of this article noted that two presentations at recent conferences concerned as yet unpublished accounts on clinical trials on deep brain stimulation that were discontinued owing to no difference between active versus sham treatment in depressed subjects. Be that as it may, at least 10 additional clinical trials on deep brain stimulation in depressed subjects are currently underway, according to records available at clinicaltrials.gov. Perhaps the outcome of those studies, once published, can provide further insight concerning the ultimate value of sham versus active deep brain stimulation in treatment-resistant depression.
Acknowledgements
The author thanks Chief Physicist Søren B. Hansen for helpful comments. Helen Mayberg, Donald A. Malone Jr., and Thomas E. Schlaepfer have been informed about the content of this article. The author declares no conflict of interest. This work was done during the author’s employment at Aarhus University.
Open access
