Vagus nerve stimulation for depression

Up to 20% of patients with depression fail to respond sufficiently to standard treatments and comprise a group who require other, and preferably better, therapies than currently available. Over the past 15 years there has been increasing therapeutic interest in using focal brain stimulation techniques for treatment-resistant depression, the rationale being that one could modulate neuronal circuitry implicated in mood and its regulation (Marangell et al. Reference Marangell, Martinez, Jurdi and Zboyan2007). One such neuromodulation approach has been vagus nerve stimulation (VNS). VNS is used to reduce seizure frequency in partial epilepsy and the observation that this seemed to improve mood in some patients under uncontrolled conditions led to its proposed use for depression.

It involves using a cardiac pacemaker-sized generator [manufactured by Cyberonics Inc., Houston, TX, USA (www.vnstherapy.com)] that is surgically implanted in the chest wall and connected to wires wrapped around the left vagus nerve. About 80% of vagus fibres are afferent and thus provide direct access to the nucleus tractus solitarius and indirectly other related deep brain structures. The VNS device costs about $15000 and the surgery takes less than an hour to complete. A hand-held telemetric device is then used to programme intermittent, but ongoing, stimulation of the vagus nerve, e.g. for 30 s every 5 min continuing daily until the stimulator is switched off. Its mechanism of action is unknown but VNS has been reported to alter metabolism of various limbic structures and levels of central neurotransmitters, providing some theoretical basis for a potential antidepressant effect (Marangell et al. Reference Marangell, Martinez, Jurdi and Zboyan2007).

Only one randomized sham-controlled trial of VNS for treatment-resistant depression (n=235) has been performed and this found no significant difference between real or sham VNS after 10 weeks of active treatment (Rush et al. Reference Rush, Marangell, Sackeim, George, Brannan, Davis, Howland, Kling, Rittberg, Burke, Rapaport, Zajecka, Nierenberg, Husain, Ginsberg and Cooke2005a). However, uncontrolled follow-up studies of 202 participants in this trial reported that after 1 year of VNS the rate for response (⩾50% reduction in baseline HRSD-24 score) was 27·2% and that after a second year of VNS about three quarters of the latter group continued to be responders (Rush et al. Reference Rush, Sackeim, Marangell, George, Brannan, Davis, Lavori, Howland, Kling, Rittberg, Carpenter, Ninan, Moreno, Schwartz, Conway, Burke and Barry2005b; Sackeim et al. Reference Sackeim, Brannan, Rush, George, Marangell and Allen2007). After 1 year of VNS the rate for remission (HRSD-24⩽9) was 15·8%. These findings raised the possibility that long-term treatment with VNS may be beneficial to a proportion of this highly resistant group. Based mainly on some of these reports, and somewhat controversially, the Food and Drug Administration (FDA) in the USA approved the Cyberonics VNS device for patients who had not responded to at least four standard treatments for depression (Lurie & Stine, Reference Lurie and Stine2006; Shuchman, Reference Shuchman2007).

Despite this FDA approval, private health insurers in the USA have been reluctant to provide routine cover for VNS for depression. In May 2007 the Centers for Medicare and Medicaid Services, national providers of health insurance in the USA, decided not to cover VNS on the grounds that ‘there is sufficient evidence to conclude that vagus nerve stimulation is not reasonable and necessary for treatment of resistant depression’ (Centers for Medicare & Medicaid Services, 2007).

In this issue, Thomas Schlaepfer and colleagues present the findings of a multi-centre open study (n=74) of 1 year of VNS in treatment-resistant depression (Schlaepfer et al. Reference Schlaepfer, Frick, Zobel, Maier, Heuser, Bajbouj, O'Keane, Corcoran, Adolfsson, Trimble, Rau, Hoff, Padberg, Müller-Siecheneder, Audenaert, Van den Abbeele, Matthews, Christmas, Stanga and Hasdemir2008). This is a partly industry-funded European replication of an open study previously carried out in the USA (Rush et al. Reference Rush, George, Sackeim, Marangell, Husain, Giller, Nahas, Haines, Simpson and Goodman2000; Nahas et al. Reference Nahas, Marangell, Husain, Rush, Sackeim, Lisanby, Martinez and George2005). They also took the opportunity to compare the European results with the earlier USA study. The results of the two studies are broadly similar with remission rates gradually increasing over the 1 year to 33% for the European group compared to 27% for the USA group who had a greater degree of treatment resistance. As the authors acknowledge, interpretation of the study is limited by it being uncontrolled but the study does confirm that VNS is reasonably well tolerated and safe. Another, possibly reassuring, conclusion to be made is that European and American patients seem to respond quite similarly to prolonged VNS treatment. However, whether this is a therapeutic or a placebo response remains unknown. So to paraphrase the popular Las Vegas slogan, as far as VNS being a clinically useful treatment for depression is concerned, the current evidence-base suggests that what happens in the vagus stays in the vagus.

Other neuromodulation techniques

What about other brain stimulation techniques for depression such as TMS or DBS? Recent randomized controlled trials of TMS for depression have been disappointing, finding no significant differences between real and sham rTMS on the primary outcome measures the trials were originally designed to test (Herwig et al. Reference Herwig, Fallgatter, Hoppner, Eschweiler, Kron, Hajak, Padberg, Naderi-Heiden, Abler, Eichhammer, Grossheinrich, Hay, Kammer, Langguth, Laske, Plewnia, Richter, Schulz, Unterecker, Zinke, Spitzer and Schonfeldt-Lecuona2007; O'Reardon et al. Reference O'Reardon, Solvason, Janicak, Sampson, Isenberg, Nahas, McDonald, Avery, Fitzgerald, Loo, Demitrack, George and Sackeim2007; Mogg et al. Reference Mogg, Pluck, Eranti, Landau, Purvis, Brown, Curtis, Howard, Philpot and McLoughlin2007). It is also becoming evident that blinding is difficult to maintain in TMS trials and this may enhance placebo effects (Mogg et al. Reference Mogg, Pluck, Eranti, Landau, Purvis, Brown, Curtis, Howard, Philpot and McLoughlin2007). TMS for depression has been reviewed by both the National Institute for Clinical Excellence in the United Kingdom (www.nice.org.uk) and the FDA in the USA (FDA Neurological Devices Panel, 2007). Preliminary reports were unsupportive regarding routine use of TMS for depression and final recommendations are imminent.

DBS requires surgical implantation of electrodes into deep brain structures and is established for intractable symptoms in Parkinson's disease and essential tremor (Ressler & Mayberg, Reference Ressler and Mayberg2007). Only a few open trials of DBS in a handful of patients have been reported for depression employing stimulation of a variety of sites including subgenual cingulate region Cg25, inferior thalamic peduncle, ventral caudate nucleus, internal globus pallidus and nucleus accumbens. While early reports appear promising, these studies have all the limitations described above and below of being unrandomized, uncontrolled and unblinded. As with VNS, potentially optimal treatment parameters (e.g. stimulation frequency, site, dose, duration, etc.) are not known for either TMS or DBS, which is a major limitation to future trial design.

The problem of control treatments and blinding

Another major difficulty for clinical trials using brain stimulation techniques is use of an appropriate sham/control treatment. In the few trials of rTMS where success of blinding has been formally measured, up to two thirds of patients correctly identify treatment (Mogg et al. Reference Mogg, Pluck, Eranti, Landau, Purvis, Brown, Curtis, Howard, Philpot and McLoughlin2007). Blinding was not measured in the largest TMS trial to date (n=301) but 35·8% of patients receiving real TMS complained of application-site pain compared to only 3·8% of the sham-treated group (O'Reardon et al. Reference O'Reardon, Solvason, Janicak, Sampson, Isenberg, Nahas, McDonald, Avery, Fitzgerald, Loo, Demitrack, George and Sackeim2007). Maintenance of blinding was not reported in the single VNS randomized controlled trial (Rush et al. Reference Rush, Marangell, Sackeim, George, Brannan, Davis, Howland, Kling, Rittberg, Burke, Rapaport, Zajecka, Nierenberg, Husain, Ginsberg and Cooke2005a). However, voice alteration was a common adverse event reported by 68% of the VNS group compared to 38% of the sham group while increase in cough occurred in 29% of the real group compared to only 9% of the sham group. Similar findings are reported by Schlaepfer and colleagues in this issue. It is therefore unlikely that blinding was successful.

It has been argued that highly treatment-resistant patients, as involved in the above studies, are less likely to experience placebo effects. However, as can be seen from Fig. 1 in Schlaepfer et al.'s paper, patients' depression ratings began to fall during the 2 weeks post-surgery and before the VNS device was turned on (Schlaepfer et al. Reference Schlaepfer, Frick, Zobel, Maier, Heuser, Bajbouj, O'Keane, Corcoran, Adolfsson, Trimble, Rau, Hoff, Padberg, Müller-Siecheneder, Audenaert, Van den Abbeele, Matthews, Christmas, Stanga and Hasdemir2008). Thus this patient group are not immune to placebo effects and this must be accounted for. As voice alteration and cough are so common with VNS, implanting a stimulator and having no stimulation at all is not going to be a suitable control. Clearly a sham VNS condition would need to control for very common side-effects and so most likely will require some degree of stimulation. However, as therapeutic stimulation parameters are unknown, one possibility is that such a potential sham condition could itself be therapeutic!

The mother of all therapeutic neuromodulation techniques is of course electroconvulsive therapy (ECT). Although not focused in its administration, ECT continues to be the most powerful treatment available for severe depression (Eranti et al. Reference Eranti, Mogg, Pluck, Landau, Purvis, Brown, Howard, Knapp, Philpot, Rabe-Hesketh, Romeo, Rothwell, Edwards and McLoughlin2007). There is therefore great merit in pursuing brain stimulation as a therapeutic approach for depression with a view to refining technique, increasing effectiveness and reducing side-effects. Having other demonstrably effective therapeutic neuromodulation techniques would be of great clinical benefit. However, until reasonable control conditions are established, and success of blinding is routinely measured, it will be difficult to interpret trial results. These are challenges to be faced rather than ignored. In the meantime, and until proven otherwise, vagus rules will continue to apply.