Other Treatment Options

Sleep Deprivation
Sleep deprivation to treat depression was first proposed by Schulte in 1969. Total sleep deprivation is defined as lack of sleep for an entire night, whereas partial sleep deprivation involves lack of sleep for half of the night. In both cases, patients do not sleep until the evening of the following day. Currently, sleep deprivation is considered by many researchers to be a safe and effective somatic treatment for depression (Pflug and Tolle 1971). Sleep deprivation has been studied as an adjunct to ongoing drug therapy (Dessauer et al. 1985; King 1980), as a switching strategy after the failure of an antidepressant (Bhanji et al. 1978; van Scheyen 1977), and as a first-line treatment for chronic depression (Svendsen 1976). In a review of 10 studies with a total of 501 subjects, the rate of response to total or partial sleep deprivation was 33%-50% (Roy and Bhanji 1976). More recent studies reported a response rate as high as 70% (Naylor et al. 1993; Vollmann and Berger 1993). However, these studies are limited by the absence of a plausible control “sham sleep deprivation” procedure (Wehr et al. 1985).

Nevertheless, most researchers using this procedure report rapid onset of improvement in patients with endogenous depression (unipolar, bipolar, psychotic, and schizophrenic depression), with the depression responding within the day after sleep deprivation (Fahndrich 1981; Larsen et al. 1976). However, the effect seems to be of limited duration, as the patients’ depression often returns after a normal period of sleep (Philipp and Werner 1979; Southmayd et al. 1990). Although early reports featured total sleep deprivation data, more recent studies indicated that partial sleep deprivation in the latter half of the night is as effective as total deprivation (Sack et al. 1988; Schilgen and Tolle 1980). Partial sleep deprivation may carry the additional advantage of avoiding the return of depressive symptoms because it may be employed more often than total deprivation, usually in a pattern of 2 nights of partial deprivation followed by 1 night of uninterrupted sleep, to sustain the response (Wehr et al. 1985). Although the antidepressant mechanism of sleep deprivation is unknown, one hypothesis is that it may have stimulant-like effects (Ebert and Berger 1998). As a low-tech solution, sleep deprivation is ripe for further study as a potential treatment for refractory depression.

Electroconvulsive Therapy
Use of ECT may be appropriate at any time during the course of unipolar illness. Although it is generally regarded as the most potent treatment for depression, limited patient acceptance and administrative obstacles result in underutilization of ECT. In view of its safety and efficacy, ECT is a particularly attractive option for patients with treatment-resistant affective illness and should be considered an early treatment option rather than one of last resort. For some patient groups—including those who are acutely suicidal, psychotic, medically complicated, or pregnant—ECT may be the first treatment of choice.

A course of ECT for acute depression typically consists of 4-12 seizures induced electrically while the patient is under anesthesia. Seizure induction takes place two to three times per week. Within the range of common clinical practice, the intensity of the electrical stimulus and the waveform (sine wave, brief pulse) seem to contribute more to differential adverse effects than to differences in efficacy.

The format of stimulus placement (bilateral vs. unilateral) appears to be associated with differences in both adverse effects and efficacy. Some studies demonstrate the antidepressant action of unilateral nondominant ECT to be the same as that of bilateral ECT, but other investigators (Mukherjee et al. 1994) and many clinicians find better response to bilateral treatment. Unilateral nondominant ECT is associated with less memory impairment than is bilateral ECT. Therefore, many clinicians will start treatment with unilateral ECT and switch to bilateral ECT only if the patient does not respond. Patients should not be considered nonresponsive to ECT unless they have had an adequate course of bilateral ECT.

Studies of ECT in depressed patients reveal a lower rate of response among patients who had not responded to antidepressant medications compared with treatment-naive patients. The Medical Research Council conducted a two-phase study on this phenomenon (Prudic et al. 1990). In the first phase, depressed patients were treated with either ECT, phenelzine, or imipramine. Nonresponders to the pharmacotherapies then received ECT. Results indicated a 71% response rate to ECT in the first phase, whereas in the second phase the response rates were significantly lower, with 50% of phenelzine nonresponders and 55% of imipramine nonresponders responding to ECT in the second phase. Patients who had previously received adequate antidepressant pharmacotherapies but had not responded were reported to have a lower response rate and a higher relapse rate after ECT treatment than patients who had responded to a prior adequate antidepressant medication trial (Prudic et al. 1990). ECT, like all effective antidepressant therapies, can induce mania in some depressed patients.

Most patients with good response to ECT will be given pharmacotherapy for continuation and maintenance treatment. Sackeim and colleagues (1990) studied the outcome of continuation and maintenance pharmacotherapy after successful treatment of acute depression with ECT. Patients are referred to ECT either because prior adequate pharmacotherapy produced no improvement (treatment failures) or because ECT was judged to be a more appropriate treatment because of acute suicidality or medical contraindication to pharmacotherapy. This study showed that ECT responders were more likely to relapse if continuation/maintenance treatment was carried out with an antidepressant agent that had previously failed to produce acute response despite an adequate trial. However, when patients had only inadequate therapy before ECT, resumption of treatment with the previous pharmacotherapy appeared to have prophylactic benefit. These data suggest that ECT responders who did not respond to a prior adequate antidepressant trial should not be given the same antidepressant for continuation or maintenance treatment.

Repetitive Transcranial Magnetic Stimulation
Repetitive transcranial magnetic stimulation (rTMS) is a technique to induce an electrical current in the brain without causing seizures. The principle of rTMS is to use a rapidly alternating current with a handheld electromagnet (generating about 2 teslas) that will in turn induce or interrupt current (depolarization) in cortical interneurons about 2 cm below the surface of the skull. One of the advantages of rTMS is that the energy is delivered without the impedance of the skull encountered in ECT. At the higher frequencies (20 Hz), rTMS is excitatory, and at the lower frequencies (5-10 Hz) it is inhibitory. The areas of greatest interest for rTMS have been the left dorsolateral prefrontal cortex (L-DLPC) and right prefrontal cortex for high and low stimulation frequencies, respectively. Extensive research efforts are aimed at clarifying the location and parameters of rTMS that might have the greatest efficacy.

Overall, the clinical effects of the early reports on rTMS were modest at best (George et al. 1999; Reid et al. 1998), whereas later studies demonstrated either no difference from placebo (Loo et al. 1999) or a robust effect (Klein et al. 1999). As pointed out by George and colleagues (1999), negative results of rTMS are not surprising given the multiple parameters that need to be worked out and the preliminary nature of the research.

Since the advent of TMS for the evaluation of neurological diseases (circa 1980), rTMS has been found to be helpful in some, but not all, studies of treatment of refractory depression. Alvaro Pascual-Leone et al. (1996) reported that 11 (65%) of 17 patients with psychotic depression who had not responded to antidepressants went on to respond to a complex protocol using 5 days of rTMS (10 Hz) with concomitant nimodipine. The data are difficult to interpret, because patients received multiple types of rTMS in various permutations of frequencies and locations of stimuli for 5 days per month over 5 months. George et al. (1997) studied 12 patients who were not treatment-resistant in a double-blind study of active compared with sham treatment in sequence. The stimulus was applied at high frequency (20 Hz) at 80% motor threshold. The researchers found that patients who were treated with the active rTMS had a modest decrease in HAM-D scores of 5 points, whereas the sham treatment resulted in an increase of 3 points. Only 4 of the 12 patients experienced a modest 25% improvement.

Figiel et al. (1998) studied 52 patients with treatment-resistant depression with rTMS using low-frequency (10 Hz), high-energy (10% above motor threshold) stimulation for 5 seconds, 10 trains each for 5 sessions at the L-DLPC. The researchers found that 22 patients (42%) responded, with response defined as a 60% decrease in baseline HAM-D score, a final HAM-D score below 16, and a Clinical Global Impressions scale (CGI; Guy 1976) score of much improved or very much improved. Loo et al. (1999) studied 18 depressed inpatients, of whom three were bipolar and one had psychotic features. This group of depressed patients did not respond to about two adequate trials of antidepressants during the current episode. Five patients had their antidepressants withdrawn, whereas the others continued to receive the antidepressants that had failed during the administration of rTMS. Patients were exposed to 10-Hz rTMS at 10% above motor threshold in 30 trains of stimuli 5 seconds each, 30 seconds apart, applied to the L-DLPC. Patients and control subjects were given real or sham rTMS for 2 weeks and were then allowed to have open treatment for another 4 weeks. No difference was found between sham and active treatments, even though the study was powered to have a 95% chance of detecting the effect size noted by Pascual-Leone et al. (1996).

In contrast to most of the therapeutic studies of rTMS, Klein et al. (1999) used low-frequency rTMS applied to the right prefrontal cortex in a controlled study of 70 inpatients. The investigators used a train of 60 stimuli of 1 Hz with 0.1-millisecond pulse duration delivered for 1 minute followed by a 3-minute interval and another 60-train stimulus given over 10 days within a period of 2 weeks. Sham treatment was applied by placing the coil at a 45-degree angle. The investigators applied the rTMS through a circular coil that provided more diffuse stimuli than the figure-8 coil used in most of the other studies. None of these patients were treatment-resistant for the current study, and 24 patients were not taking concomitant medications. In the active-treatment group, 17 (49%) of 35 patients responded (more than 50% change in HAM-D score), and in the sham group, 8 (25%) of 32 responded (P < 0.05); similarly, the active group had 16 (46%) of 35 patients with a final HAM-D score less than 10, while in the sham group only 6 (19%) of 32 achieved this posttreatment score.

At this writing of this chapter, rTMS should be considered in its infancy, a potential promising treatment for refractory depression. Additional research is needed to clarify the role of rTMS in clinical practice.

Vagus Nerve Stimulation
Stimulation of the left cervical vagus nerve, delivered by the NeuroCybernetic Prosthesis System (Cyberonics, Inc., Houston, TX) (Schachter and Saper 1998) has been commercially available in the United States for the treatment of resistant partial-onset epileptic seizures since 1997. The observation of mood improvement in patients treated for seizure disorders, positron-emission tomographic studies revealing activation of limbic structures, and neurochemical and neuronal pathway findings set the stage for an open treatment trial of vagus nerve stimulation (VNS) in treatment-resistant depression. The investigators reported improvement in 12 (40%) of a sample of 30 major depressive disorder patients with chronicity and recurrence, combined with multiple adequate treatment failures (Rush et al. 2000). Although the initial biological effects of VNS are evident, the ultimate mechanism that alleviates symptoms is as yet undiscovered (Rosenbaum and Heninger 2000; Rush et al. 2000). Controlled trials and animal studies are in preparation (J. Weiss, personal communication, February 1999), and these will be of considerable importance for establishing the efficacy and understanding the mechanism of action of VNS.