DBS in tremor

Thalamic stimulation appears to be particularly effective in the treatment of parkinsonian tremor and ET (Koller et al., 1997; Ondo et al., 1998c; Pollak et al., 1998; Koller et al., 1999; Pahwa et al., 1999; Koller et al., 2001; Krauss et al., 2001b; Rehncrona et al., 2003) (Videos 7.1 and 7.2). In the North American Multi-Center Trial, 25 ET and 24 PD patients were followed for 1 year after implantation (Koller et al., 1997). Combined blinded tremor ratings (0–4) in ET patients randomized to “on” were 0.9 compared to 2.7 for those randomized to “off” stimulation. All subjective functional measures improved, and 9 of 29 patients (31%) had complete tremor cessation. In PD patients, “on” randomized scores were 0.6 compared to 3.2 for those who were randomized to “off.” Fifty-eight percent (14 of 24) of patients had complete tremor cessation. Subjective functional measures (UPDRS part II), however, were not significantly improved. Complications were manageable and included paresthesia, headache, disequilibrium, dystonia, and device failure. Results were similar at 1 month and 1 year after implantation. Although some loss of efficacy and device-related complications have been encountered after 2 years of follow-up, the authors concluded that unilateral DBS of the thalamus has long-term efficacy in patients with ET (Koller et al., 2001). In a multicenter European study in which 37 patients with ET were followed for a mean of 6.5 years, unilateral or bilateral VIM DBS offered long-term benefits and safety (Sydow et al., 2003). Our experience at Baylor is similar to that reported in other centers. In a blinded and open-label trial of unilateral thalamic DBS in 33 patients (14 ET and 19 PD) with severe tremor refractory to conventional therapy, ET and PD patients demonstrated an 83% and 82% reduction (P < 0.0001), respectively, in observed contralateral arm tremor (Ondo et al., 1998c). All measures of tremor, including writing samples, pouring tests, subjective functional surveys, and disability scores, significantly improved. We found that bilateral thalamic DBS is more effective than unilateral DBS in controlling bilateral appendicular and midline tremors of ET and PD, and thalamic DBS does not seem to improve meaningfully any parkinsonian symptoms other than tremor (Ondo et al., 2001a). Although unilateral VIM DBS can markedly improve midline tremor, this improvement is significantly enhanced by the bilateral procedure (Putzke et al., 2005). Several studies have found that gait and balance may be impaired in some, but not all, patients treated for ET with VIM DBS (Earhart et al., 2009). One long-term study found that bilateral VIM DBS was often associated with dysarthria, loss of balance, and incoordination (Pahwa et al., 2006). Similar findings were reported in a 6-year follow-up of 38 PD patients treated with VIM DBS in a multicenter European study (Hariz et al., 2008b). The authors suggest that unilateral VIM DBS should be reserved for elderly PD patients with predominant unilateral tremor, but STN DBS may be a better choice for other patients with tremor-dominant PD. This recommendation is supported by other studies showing benefit of STN DBS in PD patients with prominent tremor (Fraix et al., 2006; Diamond et al., 2007b). VIM DBS has been found to be associated with modest improvement, rather than tremor augmentation as was previously suggested, in ipsilateral tremor in patients with ET (Ondo et al., 2001c). Other studies have demonstrated bilateral effects of unilateral thalamic DBS (Kovacs et al., 2008). A review of long-term efficacy of VIM DBS in 39 patients (20 PD, 19 ET) showed that the benefits might be maintained for at least 6 months (Rehncrona et al., 2003). In one study, three of eight patients with PD no longer required DBS after 3–5 years because the tremor markedly improved (Kumar et al., 2003). In addition to reducing the amplitude, VIM DBS increased the frequency of ET by 0.5–2 Hz at low inertial loads, made the tremor more irregular, and reduced the tremor-electromyography coherence (Vaillancourt et al., 2003).

To compare thalamic DBS with thalamotomy, Schuurman and colleagues (2000) conducted a prospective, randomized study of 68 patients with PD, 13 with ET, and 10 with multiple sclerosis. They found that the functional status improved more in the DBS group than in the thalamotomy group, and tremor was suppressed completely or almost completely in 30 of 33 (90.9%) in the DBS group and in 27 of 34 (79.4%) in the thalamotomy group. Although one patient in the DBS group died after an intracerebral hemorrhage, DBS was associated with significantly fewer complications than was thalamotomy. Similar results were obtained on this group after a 5-year follow-up, with the outcomes still favoring the DBS group (Schuurman et al., 2008). In one study of VIM DBS, dysarthria worsened when both stimulators were turned on in three of six patients (Pahwa et al., 1999). DBS also has been found to be effective in rare patients with disabling task-specific tremors (Racette et al., 2001). In addition to improving distal tremor associated with PD and ET, VIM DBS can effectively control ET-related head tremor, which usually does not respond to conventional therapy (Koller et al., 1999). Other midline tremors, such as voice, tongue, and face tremor, also may improve with unilateral VIM DBS, although additional benefit can be achieved with contralateral surgery (Obwegeser et al., 2000; Putzke et al., 2005). In addition, VIM DBS appears to improve postural stability (Pinter et al., 1999). Furthermore, unilateral thalamic DBS for ET has been found to be cognitively safe and to improve anxiety and quality of life (QoL) in terms of ADL and psychologic well-being (Tröster et al., 1999). Several studies have demonstrated that VIM DBS improves ADL and QoL, and one study showed that these improvements may be sustained for up to 7 years after implantation (Hariz et al., 2008a).

Several studies have sought to determine the optimal stimulation parameters. Algorithms for programming of the stimulation parameters have been suggested (Volkmann et al., 2006) but programming is generally based on the judgment of the programmer. Charge is the product of pulse width and current (current = voltage/impedance). Charge density is defined as the charge divided by the geometric surface area of the electrode. Short pulse widths require high current but low charge, and are thus more efficient at exciting neural elements and reducing the risk of neural damage. Linear voltage increases above 3 V do not correlate to a linear increase in the volume of neural elements excited, but significantly increase power consumption (Kuncel and Grill, 2002). While the amount of tremor suppression does not substantially increase in frequencies above 100 Hz, power consumption increased markedly at the higher frequencies (Kuncel et al., 2006). Power consumption (and battery drain) is determined by calculating the voltage squared (V²) times the pulse width (PW) and frequency (F) divided by the impedance (Z).

Bin-Mahfoodh et al. (2003) evaluated 14 implantable pulse generators (IPGs) that needed battery replacement out of 163 single channel batteries implanted over 7 years and found that the higher the total electrical energy delivered as calculated by the power consumption formula, the shorter the battery lifespan. A review of all battery replacements in patients at Baylor College of Medicine found 122 batteries needed replacement in 73 patients. The main predictors of a shorter battery life were greater amplitude, pulse width, and not using exclusive bipolar settings (Ondo et al., 2007). O’Suilleabhain et al. (2003), studied 13 VIM thalamic stimulators (8 for ET and 5 for PD) and measured tremor with an electromagnetic tracker. They randomly programmed 78 combinations of voltages (0, 1, 2, 3, or 4 V), pulse widths (60, 90, or 120 µs), and frequencies (130, 160, or 185 Hz) for both monopolar and bipolar polarity configurations. The voltage response curve for ET was flatter than for PD patients. Monopolar configurations were up to 25% more effective than bipolar but this configuration tended to shorten calculated battery life by 35%. The longest pulse width tested was up to 30% more effective than the shortest, but frequency changes had little effect on tremor amplitude. The authors recommended that there was no need to increase the frequency above 130 Hz and another study concluded that stimulation frequency of 100 Hz is generally the optimal frequency and that there is little if any additional benefit from a higher frequency (Ushe et al., 2006). Low-frequency (10 Hz) STN DBS has been shown to improve verbal fluency but worsen motor function in a double-blind randomized crossover study of 12 patients with PD (Wojtecki et al., 2006). Moro et al. (2006) studied 44 patients with PD who had a long-term (mean, 3.5 years), stable response to STN DBS. Each underwent reprogramming of their stimulation by a neurologist expert in both PD and DBS, along with medication adjustments. Scores on the UPDRS were compared before reprogramming and up to 14 months thereafter. In half the patients (24, or 54.6%), UPDRS parts II and III scores improved significantly after reprogramming, by 15.0% and 29.5%, respectively. Anti-PD drug use decreased significantly as well, by 29.5%, in these responders. No improvement was seen in 16 patients (36.4%), and the condition of 4 patients (9.1%) worsened. These results suggest that patient outcomes with STN DBS can be improved when a neurologist expert in movement disorders and DBS programming is directly responsible for patient care.

VIM DBS has been found to be useful not only in the treatment of troublesome tremors associated with ET and PD but also in orthostatic tremor (Espay et al., 2008), and cerebellar outflow tremor associated with multiple sclerosis and post-traumatic tremors (Montgomery et al., 1999; Matsumoto et al., 2001; Wishart et al., 2003; Lyons and Pahwa, 2008) (Video 7.3). Thalamic DBS may be also effective in the treatment of levodopa-induced dyskinesia, possibly owing to involvement of the center median and parafascicularis complex (Caparros-Lefebvre et al., 1999). Thalamic DBS, however, does not seem to provide any benefit in PD-related gait difficulty (Defebvre et al., 1996). Bilateral ventralis oralis anterior thalamic DBS has been also reported to be effective in a patient with severe postanoxic generalized dystonia and bilateral necrosis of the basal ganglia (Ghika et al., 2002). Medically intractable myoclonus was reported to improve 80% with VIM DBS in one case of myoclonus–dystonia syndrome (Trottenberg et al., 2001a), and GPi DBS has been found to be effective in relieving myoclonus–dystonia syndrome (Magarinos-Ascone et al., 2005).

In addition to VIM as a target for patients with disabling tremor, STN DBS has been also found to be useful in the treatment of severe proximal tremor (Kitagawa et al., 2000) as well as rest and postural tremor in PD (Sturman et al., 2004).

DBS in Parkinson disease and levodopa-related complications

Several studies have demonstrated that DBS of the GPi and STN improves not only parkinsonian tremor but also other PD-related symptoms and prolongs the “on” time (Pahwa et al., 1997; Limousin et al., 1998; Pollak et al., 1998; Moro et al., 1999; Deep-Brain Stimulation for Parkinson’s Disease Study Group, 2001; Lanotte et al., 2002; Alvarez et al., 2005) (Videos 7.4, 7.5). While levodopa improves predominantly finger bradykinesia, STN DBS improves predominantly proximal and axial bradykinesia (Timmermann et al., 2008). Elderly patients (>70 years old) seem to benefit as much as younger patients with respect to a reduction in dyskinesias and motor fluctuations, but during the “off” times, older patients have more difficulties with ADL and axial symptoms, particularly if they had some gait problems before surgery (Russmann et al., 2004b). Depending on the location of the stimulating electrode, pallidal stimulation has a variable effect on parkinsonian features versus levodopa-induced dyskinesias. In one study, stimulation of the dorsal GPi improved parkinsonian features, but stimulation of the posteroventral GPi improved levodopa-induced dyskinesia and worsened gait and akinesia (Bejjani et al., 1997). In another study, stimulation of the most ventral part of the GPi improved rigidity and eliminated levodopa-induced dyskinesia but produced marked bradykinesia, whereas stimulation of the most dorsal contacts improved bradykinesia and induced dyskinesia (Krack et al., 1998a). Best results could be obtained by stimulating the intermediate contacts. Similar results were reported by Durif and colleagues (1999), who found that ventral GPi stimulation was more effective than dorsal stimulation for alleviating rigidity and levodopa-induced dyskinesia. But in contrast to Krack and colleagues (1998a), whose target was posterolateral to the location of the target used by Durif and colleagues (1999), the latter group found that ventral stimulation also improved bradykinesia. They concluded that “chronic stimulation in the anteromedial GPi shows that this is a safe and effective treatment for advanced PD.” GPi DBS has been found not only to improve “off” state UPDRS and movement onset time and spatial errors, but also to enhance motor activation responses as measured by concurrent PET recordings of regional cerebral blood flow in the sensorimotor cortex, supplementary motor area, and anterior cingulated gyrus (Fukuda et al., 2001a). Normalization of abnormal pattern of cerebral blood flow and an increase in cerebral blood flow in the supplementary motor area and anterior cingulated cortex has been also found with STN DBS, and this correlated with improvement in bradykinesia (Strafella et al., 2003). This effect was more robust with bilateral than unilateral stimulation. Hershey and colleagues (2003), however, found that STN DBS increased blood flow in the midbrain, globus pallidus, and thalamus but decreased blood flow in cortical areas. They concluded that “STN stimulation appears to drive, rather than inhibit, STN output neurons.” This is consistent with the model that increased STN output driven by STN DBS leads to excitation of pallidal neurons, which increases inhibitory output to the thalamus resulting in thalamocortical inhibition. This is supported by the observed decrease in overactivity of SMA and other areas of overactivity presumably recruited to compensate for abnormal motor initiation as measured by resting regional cerebral blood flow (rCBF) (Grafton et al., 2006). In one study, bilateral STN DBS was associated with increased rCBF in both thalami and right midbrain and with decreased rCBF in the right premotor cortex (Karimi et al., 2008). The increased rCBF in the thalami correlated with improved bradykinesia and decreased rCBF in the SMA correlated with improvement in rigidity. It is, however, still not fully understood how rCBF reflects “normalization” of the abnormal pattern of neuronal firing in the target nuclei.

The localization of STN has been facilitated by the neurophysiologic characterization of the STN (Hutchison et al., 1998). The mean firing rate was found to be 37 Hz. The firing pattern was irregular and movement-sensitive. In addition, tremor cells were identified in the STN and ventral pallidum. Macroelectrode STN stimulation completely suppressed contralateral tremor (Ashby et al., 1999). In a 1-year follow-up of 24 PD patients treated with bilateral STN stimulation, the UPDRS, parts II and III, improved by 60%, and the mean dose of dopaminergic drugs was reduced by half (Limousin et al., 1998). Krack and colleagues (1998b) found a 71% reduction in the UPDRS score in patients treated with STN DBS. Using subjective evaluation of contralateral wrist rigidity, Rizzone and colleagues (2001) studied the effects of various STN stimulation parameters in patients with PD during a drug-free state. They found that stimulus rates higher than 90 Hz and a pulse width larger than 60 µs rarely produce additional benefits and are associated with narrowing of the therapeutic window. On the basis of longitudinal experience in a large number of patients, a monopolar stimulation with 2 V, pulse width 60 µs, and frequency of 100–130 Hz is usually found to provide maximal benefit (Ushe et al., 2004). There is a characteristic pattern of emergence of tremor within minutes of discontinuation of the DBS, followed by slow and steady worsening of axial signs over 3–4 hours, with 90% of the UPDRS score worsening occurring within 2 hours after DBS is turned off (Temperli et al., 2003). When the STN DBS is turned on, a similar but faster rate of improvement is observed. In an analysis of the effects of 25 electrodes used in STN DBS, placement in the dorsolateral STN border was associated with best clinical results and least energy consumption (Herzog et al., 2004) and stimulation of the dorsal border of STN has been associated with the most robust clinical improvement (Maks et al., 2009). Furthermore, ventral stimulation was associated with more adverse cognitive and mood changes (Okun et al., 2009).

Stereotactic surgery is the most effective and predictable treatment for levodopa-induced dyskinesia resistant to pharmacologic therapy (Guridi et al., 2008). In a multicenter, prospective, double-blind, crossover study in 134 patients with advanced PD treated with DBS of STN or GPi, the UPDRS motor scores improved by 49% (P < 0.001) and 37% (P < 0.001), respectively, in comparison to the nonstimulated state (Deep-Brain Stimulation for Parkinson’s Disease Study Group, 2001). Furthermore, 6 months following implantation as compared to baseline, the percent time “on” without dyskinesias increased from 27% to 74% (P < 0.001) and from 28% to 64% (P < 0.001) with STN and GPi DBS, respectively. While the levodopa dosage remained unchanged in the GPi group, the daily levodopa dose equivalents were reduced by 37% in the STN DBS group (P < 0.001). Adverse events included intracranial hemorrhage in seven patients and infection necessitating removal of the leads in two. Although the largest and best-designed study, it was criticized because of methodologic flaws (e.g., absence of true blindness), short follow-up, and underreporting of adverse effects (Obeso et al., 2002). In a blinded assessment at 1 year after implantation of STN DBS in 30 patients, the motor UPDRS decreased by only 30%, but duration of daily wearing-off decreased by 69%, and levodopa requirements decreased by 30% (Ford et al., 2004). In a double-blind study of 7 patients treated with STN DBS, Kumar and colleagues (1998a) found 58% improvement in the UPDRS motor score in a medication-off state when the stimulator was turned on. Furthermore, there was an 83% improvement in levodopa-induced dyskinesias, and the total drug dosage decreased by 40%. In another study involving 23 patients with PD treated with bilateral STN DBS, Houeto and colleagues (2000) showed that the procedure decreased levodopa-induced motor fluctuations, dyskinesias, and daily dose of levodopa by 61–78%. Similarly, Fraix and colleagues (2000) showed in 24 patients with PD treated with bilateral STN DBS that the observed improvement in levodopa-induced dyskinesia was associated with a reduction in levodopa dosage. In another study, the combination of reduced levodopa and STN stimulation reduced the duration of diphasic and peak-dose dyskinesias by 52% and reduced “off” period dystonia by 90% and “off” period pain by 66% (Krack et al., 1999). In another, much smaller study involving only 15 patients with PD treated with bilateral STN DBS, the overall levodopa daily dose was reduced by 80.4%, and levodopa was withdrawn in 8 (53%) patients (Molinuevo et al., 2000). In a long-term follow-up, they showed that 10 of 26 (38%) patients were maintained on STN DBS monotherapy after 1.5 years of treatment (Valldeoriola et al., 2002). Similarly, Vingerhoets and colleagues (2002) found that 21 ± 8 months after implantation of bilateral STN DBS under stereotactic guidance, microelectrode recording, and clinical control in 20 patients, the UPDRS III “off medication” score decreased by 45% and was similar to UPDRS III “on medication” score. Furthermore, medication was reduced by 79%, and 10 (50%) were able to withdraw their medications completely. In a subsequent study of two groups of six STN DBS-treated patients, the investigators showed that patients who were able to discontinue their medication and were subsequently challenged with levodopa had much less severe levodopa-induced dyskinesia than those who had continued on levodopa, thus supporting “dopaminergic stimulation and striatal desensitization as major determinants of levodopa-induced dyskinesia in PD” (Russmann et al., 2004a). In a 2-year follow-up of 20 patients with STN DBS, the 50.9% UPDRS motor score reduction observed at 6 months was maintained during the follow-up period (Herzog et al., 2003). In a 5-year prospective study of the first 49 patients, mean age 55 years, treated with bilateral STN DBS, assessed during “on” and “off” states at 1, 3, and 5 years, the Grenoble group found a 54% improvement in “off” motor function compared to baseline and 49% improvement in ADL (Krack et al., 2003). Speech was apparently the only motor function that did not improve. Except for improved dyskinesia and lower daily levodopa dose, there was no additional improvement in “on” motor function beyond 1 year, and the axial symptoms continued to deteriorate after the first year. Of the initial 49 patients, 7 did not complete the study, 3 died, 4 were lost to follow-up, 3 developed dementia after 3 years, 1 committed suicide, and 1 had a large cerebral hemorrhage. This and other studies provide evidence for the conclusion that STN DBS is no better than levodopa, but it ameliorates levodopa-related motor complications and dyskinesias and “off” period dystonia. In a double-blind, crossover evaluation of STN DBS in 10 patients during practically defined “off” (medications stopped overnight), the mean motor UPDRS score improved from 43 to 26 (P < 0.04), and various timed tests (walking and tapping) also improved significantly (P < 0.04) (Rodriguez-Oroz et al., 2004). Open-label, 4-year follow-up also showed significant improvement in dyskinesia and levodopa reduction by half. In a 5-year follow-up of 11 patients treated with bilateral pallidal DBS, dyskinesias remained well controlled, but the initial benefit in “off” motor symptoms and fluctuations as well as ADL gradually declined, and the benefits were restored in 4 patients after replacement of the pallidal electrodes into STN (Volkmann et al., 2004). These results are consistent with many other studies showing long-term benefits of STN DBS in PD (Wider et al., 2008). In a study of 30 patients undergoing STN DBS, levodopa-induced complications based on UPDRS IV and during an acute levodopa challenge improved markedly after 1 year (both on and off stimulation) and still further at 5 years (Simonin et al., 2009). Furthermore, the peak-dose dyskinesia decreased significantly between 1 and 5 years, possibly related to a reduction in levodopa-equivalent dose from 1323 ± 501 mg at baseline to 753 ± 451 mg at year 1, and 850 ± 555 mg at year 5. The DBS parameters at 5-year follow-up were as follows: voltage 2.92 ± 0.73 V, frequency 127 ± 28 Hz and pulse 76 ± 14 µs. Reduction of dopaminergic drugs as a result of STN DBS may unmask restless legs syndrome (Kedia et al., 2004). STN DBS, however, has been reported to improve restless legs syndrome in 6 patients in whom the syndrome accompanied PD (Driver-Dunckley et al., 2006). Bilateral STN DBS has been reported to produce about 78% improvement in the thoracolumbar angle of patients with PD-associated camptocormia (Sako et al., 2009).

Besides reducing the need for levodopa, bilateral STN DBS also reduces the need for other medications, including apomorphine (Varma et al., 2003). Moro and colleagues (2002) showed that while the duration and latency of levodopa response are well maintained in patients with chronic STN DBS, the magnitude of the short-duration response tends to decrease with time. Since improvements in dyskinesia usually require a reduction in levodopa dosage, unilateral STN DBS is impractical because the side of the body contralateral to the unstimulated side would clearly worsen. Many investigators recommend bilateral STN DBS even in markedly asymmetric PD (Kim et al., 2009). Furthermore, bilateral implantation during a single procedure is less inconvenient to the patient than a staged procedure, the neurophysiologic mapping may be facilitated by anatomic symmetry, and implantation of both stimulators can be performed under a single general anesthetic.

Bilateral STN DBS appears to be more effective than unilateral STN DBS in improving parkinsonism, but unilateral STN DBS may be appropriate for patients with asymmetric parkinsonian symptoms, including a high-amplitude tremor (Kumar et al., 1999b; Linazasoro et al., 2003; Sturman et al., 2004; Stover et al., 2005; Diamond et al., 2007b; Fishman, 2008). STN DBS has been shown to have a robust effect not only on levodopa-related motor complications but also on the treatment of PD-related tremor (Diamond et al., 2007b). This marked benefit on tremor is also supported by marked improvement of postural and kinetic tremor with stimulation of the subthalamic area (Hamel et al., 2007; Herzog et al., 2007). Since bilateral STN DBS appears to be associated with less ataxia and dysarthria than VIM DBS (Pahwa et al., 2006), it may be considered as the target of choice in patients with ET requiring bilateral stimulation (Diamond et al., 2007b).

Bilateral STN DBS greatly improves functioning and reduces levodopa-induced dyskinesias, probably by allowing a reduction in total levodopa dosage. STN DBS, however, may also somehow reverse the levodopa sensitization, since levodopa-induced dyskinesias seem to be markedly reduced after continuous bilateral STN stimulation even when the DBS is turned off (Bejjani et al., 2000b). This may be due to the marked reduction in daily levodopa dose permitted by chronic STN DBS. Nutt and colleagues (2001b) suggested that the improvement in motor fluctuations produced by STN and GPi DBS is due to improvement in “off” disability rather than any effects on pharmacodynamics or pharmacokinetics of levodopa. Younger patients with levodopa-responsive PD are considered the best candidates for bilateral STN DBS (Charles et al., 2002; Welter et al., 2002). In a 2-year follow-up of patients with bilateral STN DBS, preoperative response to levodopa was found as the only predictor of a favorable outcome (Kleiner-Fisman et al., 2003). In patients with prior pallidotomy, bilateral (not unilateral) STN DBS may be beneficial (Su and Tseng, 2001; Kleiner-Fisman et al., 2004). The mean firing frequency of STN on the side ipsilateral to the pallidotomy is lower than on the contralateral, intact side (Mogilner et al., 2002). In a randomized trial involving 34 patients with advanced PD the “off” UPDRS score improved significantly more following bilateral STN DBS as compared to unilateral pallidotomy; “on” UPDRS motor and dyskinesia scores also improved more in the DBS group than in the pallidotomy group (Esselink et al., 2004).

Although bilateral STN DBS is clearly effective in improving the cardinal as well as other parkinsonian symptoms, the procedure does not necessarily improve all the symptoms, and as a result of progression of cognitive, speech, and other deficits frequently associated with PD, the QoL might not substantially improve in patients who exhibit these additional features (Hariz et al., 2000; Diamond and Jankovic, 2005), only about half of patients regain their employment, and many experience new marital and socio-professional problems (Schupbach et al., 2006). In one study, only young PD patients were found to have a significant improvement in their quality of life with STN DBS (Derost et al., 2007). Furthermore, preexisting personality disorders and psychiatric disorders might not necessarily improve with STN DBS and might actually worsen and further compromise the patient’s quality of life (Houeto et al., 2002), but in carefully selected patients, mood, anxiety, apathy, and quality of life may improve with the procedure (Czernecki et al., 2005; Houeto et al., 2006). Normalization of bladder symptoms associated with PD-related detrusor hyperreflexia was demonstrated in a study of 16 patients with bilateral STN DBS (Seif et al., 2004).

Using the PD Quality of Life Questionnaire in 60 patients, Lagrange and colleagues (2002) showed improvements 1 year after bilateral STN DBS in all dimensions, including motor (+48%), systemic (+34%), emotional (+29%), and social (+63%). Other studies have demonstrated significant improvements in health-related quality of life in patients with advanced PD treated with bilateral STN DBS (Just and Ostergaard, 2002; Lagrange et al., 2002; Martinez-Martin et al., 2002). In another study, Lezcano and colleagues (2004a) applied the UPDRS, PDQ-39, and the scale of quality of life for caregivers (SQLC) in 11 PD patients 2 years after they had undergone bilateral STN DBS and found 62% improvement in PDQ-39 (P < 0.001) and 68% in SQLC (P = 0.002). In an 18-month follow-up of patients after bilateral implantation of STN DBS, sustained improvements were demonstrated in a variety of measures of health-related quality of life (Siderowf et al., 2006). A cost-effectiveness analysis suggests that DBS could be cost-effective in treating PD quality of life, which improved 18% or more compared with best medical treatment (Tomaszewski and Holloway, 2001). In a European, SPARK, study, the 6-month cost of PD decreased from about $10 000 to $1600 after bilateral STN DBS, largely due to marked reduction in medications (Fraix et al., 2006). In a 6-month study of 156 patients with advanced PD randomized to STN DBS or medical management, the German Parkinson Study Group (Deuschl et al., 2006b) concluded that STN DBS was more effective than medical management. The baseline characteristics, including the mean age of the patients in each group (60.5 vs. 60.8 years, respectively) was similar, but the mean PDQ-39 score, the primary endpoint, decreased from 41.8 to 31.8 in the DBS group (25% improvement) and increased from 39.6 to 40.2 in the medication group (no improvement). The PDQ-39 improvement in the DBS group was significant in the subscales for mobility, ADL, emotional well-being, stigma, and bodily discomfort. The mean UPDRS-III score during “off” time improved from 48.0 to 28.3 (41% improvement) with DBS, but remained unchanged in the medication group (46.8 vs. 46.0) and improved during “on” time from 18.9 to 14.6 (23% improvement) in the DBS group, but remained unchanged in the medication group (17.3 vs. 17.5). Although a total of 39 (50%) adverse events were reported in the DBS group as compared to 50 (64%) in the medication group, there were significantly more serious adverse events reported in the DBS group (n = 10, 12.8%), including 3 deaths, as compared to the medication group (n = 3, 3.8%) (P = 0.04). In another prospective study of 20 patients with early PD, randomized to either STN DBS or medical therapy, the authors found that the quality of life measures improved 24% in the surgical group but not at all in the medical group and after 18 months the severity of parkinsonian motor signs “off” medication, levodopa-induced motor complications, and daily levodopa dose were reduced by 69%, 83%, and 57% in operated patients and increased by 29%, 15%, and 12% in the group with medical treatment only (P < 0.001) (Schupbach et al., 2006). The authors concluded that DBS “should be considered a therapeutic option early in the course of PD.” A total of 255 patients were enrolled in a randomized, controlled trial, designed to compare the effects of DBS (STN, n = 60 or GPi, n = 61) and “best medical therapy” (n = 134) after 6 months, at seven Veterans Affairs (The Parkinson’s Disease Research, Education, and Clinic Center, PADRECC) and six university hospitals (Weaver et al., 2009). Patients treated with DBS gained a mean of 4.6 hours/day of “on” time without troubling dyskinesia compared with 0 hours/day for patients who received best medical therapy (P < 0.001). Furthermore, motor function improved by ≥5 points on the motor UPDRS in 71% of DBS and 32% of medical therapy patients. This was accompanied by improvements in the majority of PD-related HRQoL measures and only minimal decrement in neurocognitive testing. The overall risk of experiencing a serious adverse effect, however, was 3.8 times higher in the DBS than in the medical therapy group (40% vs. 11%). While benefits associated with DBS extend beyond what can be achieved with medical therapy alone, selection of the appropriate patients and target as well as skills and experience of the DBS team must be considered before referring patients for this surgical treatment (Okun and Foote, 2009). It is also important when selecting candidates for surgery to carefully evaluate all the ethical aspects of DBS, the net benefit, and the long-term burden it may place on the patient and his or her caregiver, especially given the impaired cognition that can follow the implantation of the DBS (Farris et al., 2008a). Although there are many reports concluding that GPi DBS is less effective than STN stimulation, there are only few studies that have objectively compared these two approaches (Okun and Foote, 2005; Okun et al., 2009) (Fig. 7.4). Before addressing this controversy, it is important to point out that GPi is a much larger target than STN (500 mm³ versus 200 mm³) and therefore requires larger density of stimulation. On the other hand, stimulation of the STN, because it is a smaller target, may lead to spreading of current from the intended sensorimotor area to potentially unwanted areas of the nucleus, such as the associative and limbic regions. This might account for a slightly higher frequency of cognitive and behavioral adverse effects with STN DBS (Anderson et al., 2005). Several studies have demonstrated greater antidyskinetic effects of GPi versus STN stimulation, and some studies have suggested that GPi DBS might be especially indicated in patients with a low threshold for dyskinesia (Minguez-Castellanos et al., 2005), and GPi is considered the desirable target for treatment of dystonia and other hyperkinetic movement disorders. The relative safety and efficacy of STN versus GPi DBS have been compared in a few studies (Burchiel et al., 1999; Anderson et al., 2005; Okun et al., 2009). Although Burchiel and colleagues (1999) found no difference between the two targets, their initial study had insufficient power to detect a difference between the two groups. In their subsequent study involving 23 patients with PD complicated by marked levodopa-related motor fluctuations and dyskinesias, the investigators concluded that there were no significant differences in the overall benefits, although levodopa was decreased more in the STN group and dyskinesia improved more in the GPi group (Anderson et al., 2005). Krause and colleagues (2001) reported results in 6 patients with PD treated with GPi DBS and 12 patients with STN DBS. They found that while GPi DBS was associated with a lower frequency of levodopa-related complications, there was no improvement in bradykinesia or tremor. STN DBS, on the other hand, improved all parkinsonian symptoms and was associated with a reduced daily dose of levodopa. The authors concluded that STN was “the target of choice” for patients with severe PD who have side effects from levodopa. In a retrospective analysis of 16 patients undergoing STN DBS versus 11 treated with GPi DBS, Volkmann and colleagues (2001) showed that STN stimulation was associated with a 65% reduction in medication and required less electric power but required more intensive postoperative monitoring and was associated with a higher frequency of adverse effects related to levodopa withdrawal as compared to GPi DBS. When STN and GPi DBS were compared, no difference between stimulation of the two targets on improvement of rigidity, strength, speed of movement execution, or movement initiation was found (Brown et al., 1999). For most effective results, the upper STN (sensorimotor part) and the subthalamic area containing the zona incerta, fields of Forel, and STN projections should be stimulated (Hamel et al., 2003). However, further controlled, randomized studies are needed to answer the question of which of the two methods is more effective and which patients should be considered the best candidates for either of the two procedures. In the COMPARE trial 52 patients were randomized to unilateral STN or GPi DBS without any significant difference in outcomes except more worsening in letter verbal fluency in patients with STN DBS (Okun et al., 2009). In the a follow-up analysis of the Veterans Affairs Cooperative Studies Program (Weaver et al., 2009) outcomes, STN and GPi DBS were analyzed after 24 months in 299 patients; there were no differences in mean changes in the motor (Part III) UPDRS between the two targets (Follett et al., 2010). Patients undergoing STN required a lower dose of dopaminergic agents than those undergoing pallidal stimulation (P = 0.02) and visuomotor processing speed declined more after STN than after GPi stimulation (P = 0.03). On the other hand, there was worsening of depression after STN DBS but mood improved after GPi DBS (P = 0.02). Slightly more than half of the patients experienced serious adverse events but there was no difference in the frequency of these events between the two groups. As the results of comparative trials become available there is emerging evidence that GPi DBS may be particularly suitable for patients who may have some cognitive or behavioral issues whereas bilateral STN DBS may be the surgical choice for patients in whom reduction in levodopa dosage is the primary goal.

Figure 7.4 Results of a multicenter trial of STN and GPi DBS.

Redrawn from Deep-Brain Stimulation for Parkinson’s Disease Study Group. N Engl J Med 2001;345:956–963.

DBS of the STN has been found to be effective in controlling not only parkinsonian tremor (Sturman et al., 2004), but also bradykinesia, gait difficulty, and freezing (Limousin et al., 1995; Allert et al., 2001; Faist et al., 2001; Stolze et al., 2001; Bakker et al., 2004; Ferraye et al., 2010; Moro et al., 2010) and handwriting (Siebner et al., 1999). In addition to improving limb signs and the cardinal signs of PD, bilateral STN DBS has been found to also improve axial parkinsonian symptoms, particularly rising from chair and gait (Bastian et al., 2003; Bakker et al., 2004), as well as speech, neck rigidity, abnormal posture, “off” dystonia, balance and postural instability (Maurer et al., 2003; Colnat-Coulbois et al., 2005; Nilsson et al., 2005; Shivitz et al., 2006), and sensory symptoms (Bejjani et al., 1999; Loher et al., 2002; Rocchi et al., 2002; Krystkowiak et al., 2003). Even sexual functioning has been shown to improve after STN DBS (Castelli et al., 2004). Using static posturography, Rocchi and colleagues (2002) found an improvement in postural sway with bilateral STN DBS but worsening with levodopa. Bilateral STN DBS has been reported to improve PD-related dysarthria in some studies (Gentil et al., 1999; Pollak, 1999). Other studies, however, have found that dysarthria, cognitive impairment, and postural instability fail to improve with STN DBS even in the same PD patient in whom the procedure improved the usual levodopa-responsive symptoms (Jarraya et al., 2003). STN or GPi DBS has been shown to improve gait velocity by increasing stride length with normalization of the gait pattern (Pahwa et al., 1997; Damier et al., 2001; Allert et al., 2001; Faist et al., 2001; Bastian et al., 2003). In one study, the “off” period cadence increased from 117 ± 18.9 steps per minute to 126 ± 9.4 steps per minute (P < 0.05) after GPi DBS (Volkmann et al., 1998). In contrast, bilateral STN DBS increased stride length but not cadence (Allert et al., 2001; Faist et al., 2001). Other studies reported significant improvements in essentially all components of gait (Krystkowiak et al., 2003; Bakker et al., 2004).

The possibility that chronic DBS interferes with STN’s excitatory output suggests a potential role of this treatment as a neuroprotective strategy (Piallat et al., 1996; Henderson and Dunnett, 1998; Krack et al., 1998b). This notion is supported by the observation that ablation of STN attenuates the loss of DA neurons in rats exposed to the mitochondrial toxin 3-nitroproprionic acid (3-NP) or the catecholamine toxin 6-hydroxydopamine (6-OHDA) (Nakao et al., 1999). A 20–40% preservation of SN neurons was observed with STN lesion of DBS using MPTP monkeys as a model for PD (Wallace et al., 2007). The animal studies have led some investigators to argue that STN DBS may offer neuroprotection by reducing glutamate excitotoxicity and that performing the procedure early in the course of the disease might prevent motor disability and adverse reactions to levodopa (Mesnage et al., 2002; Schupbach et al., 2007; Wallace et al., 2007).

Effects of ablative surgery versus DBS on various parkinsonian features have not been adequately compared. In one prospective study, 13 patients with PD were randomized to pallidal stimulation versus pallidal ablation (Merello et al., 1999). Although the primary endpoint effects at 3 months on UPDRS and ADL were “comparable,” pallidal DBS had more beneficial effects on hand-tapping speed, whereas pallidotomy had more robust effect on levodopa-induced dyskinesias. However, additional blinded comparison studies of STN versus GPi DBS are needed before any definite conclusions about the relative effects of the two targets can be made. DBS electrodes are increasingly being used to make ablative lesions in the target areas (Raoul et al., 2003). GPe DBS might also provide benefits to patients with PD and with less delay than GPi DBS, although in one study, it was associated with slightly higher frequency of dyskinesia (20% versus 9%) (Vitek et al., 2004).

As with any procedure, appropriate selection of patients as candidates for DBS surgery is critical to ensure optimal outcome (Table 7.5). Screening tools for surgical candidates have been developed and validated (Okun et al., 2004a; Tan and Jankovic, 2009). Many studies have demonstrated that patients with atypical parkinsonism, such as progressive supranuclear palsy, multiple system atrophy, or dementia with Lewy bodies, do not respond well to surgery and their symptoms may substantially worsen after surgery. Patients with atypical parkinsonism, such as multiple system atrophy, are poor candidates for surgery, and STN DBS may aggravate speech, swallowing, and gait (Tarsy et al., 2003; Lezcano et al., 2004b; Shih and Tarsy, 2007).

Table 7.5 Selection of patients for deep brain stimulation surgery

Since the diagnostic inaccuracy improves with follow-up, it is advisable to consider patients for surgery only when there is a considerable duration of symptoms (at least 5 years). This is reasonable considering the fact that most patients develop complications of fluctuations and dyskinesias, the chief indications for surgery, only after 5–6 years of levodopa therapy.

Besides GPi, GPe, and STN, other targets that are being explored for DBS treatment of PD include the caudal part of the zona incerta (cZI) (Plaha et al., 2006, 2008a, 2008b) and the PPN (Hamani et al., 2007; Stefani et al., 2007). Several groups have reported that the most effective target for treatment of PD lies dorsal/dorsomedial to the STN (region of the pallidofugal fibers and the rostral zona incerta) or at the junction between the dorsal border of the STN and the latter. In a study of 35 patients with PD who underwent MRI-directed implantation of 64 DBS leads into the STN (17), dorsomedial/medial to STN (20), and cZI (27), the mean adjusted contralateral UPDRS III score with cZI stimulation was 3.1 (76% reduction) compared to 4.9 (61% reduction) in the group with target dorsomedial/medial to STN, and 5.7 (55% reduction) in the STN (P value for trend <0.001) (Plaha et al., 2008a). There was a 93% improvement in tremor with cZI stimulation, but no improvement in dyskinesia scores and there was no change in levodopa dosage. The authors concluded that “High frequency stimulation of the cZI results in greater improvement in contralateral motor scores in PD patients than stimulation of the STN.” The authors also showed that rest tremor of PD can be induced by 5–40 Hz stimulation of cZI (Plaha et al., 2008a) and that PD tremor, essential tremor, and other tremors may be markedly improved with low-voltage, high-frequency DBS of cZI (Plaha et al., 2008b). A pathologically proven case of cZI DBS provided evidence for effective treatment of PD symptoms with stimulation of this target (Guehl et al., 2008).

Although low-frequency (60 Hz) as compared to the standard high-frequency (130 Hz) STN stimulation has been reported to benefit freezing (Moreau et al., 2008), this PD symptom is usually refractory to both medical and surgical therapy. High-frequency STN DBS does not appear to provide additional benefit on freezing beyond levodopa (Ferraye et al., 2010).

Although PPN DBS has been reported to provide improvement in gait in monkeys, targeting this brainstem nucleus for therapeutic purposes in patients with PD or other movement disorders associated with gait difficulty will be technically challenging (Jenkinson et al., 2006). PPN has been proposed as a potential therapeutic target in patients with gait and balance difficulties, including freezing (Nashatiazadeh and Jankovic, 2008; Strafella et al., 2008; Ferraye et al., 2010; Moro et al., 2010). Six patients with unsatisfactory pharmacologic control of axial signs such as gait and postural stability underwent bilateral implantation of DBS electrodes in the STN and PPN (Stefani et al., 2007). Clinical effects were evaluated 2–6 months after surgery in the off- and on-medication state, with both STN and PPN stimulation on or off, or with only one target being stimulated. Bilateral PPN DBS at 25 Hz in off-medication produced an immediate 45% amelioration of the motor UPDRS subscale score, followed by a decline to give a final improvement of 32% in the score after 3–6 months. In contrast, bilateral STN DBS at 130–185 Hz led to about 54% improvement. PPN DBS was most effective on gait and postural items. In on-medication state, the association of STN and PPN DBS provided a significant further improvement when compared to the specific benefit mediated by the activation of either single target. Moreover, the combined DBS of both targets promoted a substantial amelioration in the performance of daily living activities. The authors felt these findings indicate that in patients with advanced PD, PPN DBS associated with standard STN DBS may be useful in improving gait and in optimizing the dopamine-mediated on-state, particularly in those whose response to STN-only DBS has deteriorated over time and may also prove useful in disorders, such as progressive supranuclear palsy. In another study, involving 6 patients with PD, unilateral PPN DBS was reported to reduce frequency of falls (Moro et al., 2010), but further studies are needed to confirm this initial finding. Patients with significant freezing of gait were not evaluated in this study, but were the subjects of other reports. In one patient, a 63-year-old man with PD-related freezing of gait, postural instability, and marked bradykinesia, PPN DBS resulted in 20% improvement in motor function, associated with increased regional cerebral blood flow in both thalami (Strafella et al., 2008). At Baylor College of Medicine, we assessed the effects of PPN DBS in three patients. Patient 1, a 59-year-old man, underwent bilateral PPN DBS with noticeable improvement in his freezing of gait 2 and 6 weeks postoperatively, but the improvement was not subsequently sustained despite adjustments of stimulating parameters (Nashatiazadeh and Jankovic, 2008). Patient 2, a 76-year-old man, previously received unsuccessful bilateral GPi DBS, but had transient improvement in freezing after DBS of the right PPN. Patient 3, a 66-year-old man, underwent isolated right PPN but developed an increase in his Gait and Balance Scale score without noticeable improvement in freezing. In all three patients DBS setting adjustments were limited by blurry vision and balance problems. Further studies are needed before PPN DBS can be routinely recommended for the treatment of gait and balance problems and freezing of gait. In a study of 6 patients with PD-related freezing of gait PPN DBS resulted in modest improvements with overall results described as “disappointing” (Ferraye et al., 2010).

DBS safety and complications

STN stimulation, while usually well tolerated, may produce ballistic and choreic dyskinesia when the voltage is increased above a given threshold, which is probably different from that produced by levodopa (Moro et al., 1999) (Video 7.6). Except for mild deficit in verbal memory and fluency, STN or GPi DBS does not appear to adversely affect cognitive performance (Pillon et al., 2000; Heo et al., 2008). In a study at Baylor College of Medicine, when 23 patients undergoing bilateral STN DBS were compared with 28 treated medically at 6 months after surgery, the DBS patients were found to have mild but significant declines in verbal recall and fluency and mild decline in frontostriatal cognitive measures, even when good motor outcome was achieved (York et al., 2008). When followed for 2 years, the STN DBS patients demonstrated significant impairments in verbal memory, oral information processing speed, and language (Williams et al., 2010). Additionally, one-third of the STN DBS patients declined on confrontational naming compared to none of the PD patients, and one-third of the STN DBS patients progressed to PD dementia 2 years following STN DBS compared to none in the PD group. These results are consistent with the more recent reports on STN DBS that have found long-term deficits in multiple cognitive domains, including memory, attention, and frontostriatal functioning following STN DBS when more comprehensive neuropsychologic measures were administered (Contarino et al., 2007). In one controlled study 60 patients were randomly assigned to receive STN DBS and 63 to have best medical treatment. After 6 months, DBS-treated patients showed mild but significantly more evidence of impairments in executive function and verbal fluency, irrespective of the improvement in quality of life (Witt et al., 2008). In contrast, anxiety was reduced in the DBS group compared with the medication group. Based on data collected from eight advanced PD patients, mean age 56.5 years, during off-medication periods, significant declines in cognitive and motor function were found under modest dual-task conditions with bilateral but not with unilateral STN DBS (Alberts et al., 2008). Impaired performance in verbal fluency associated with STN DBS has been correlated with decreased regional cerebral blood flow in the left frontotemporal areas as measured by PET (Schroeder et al., 2003). While many studies have concluded that cognitive function is unchanged with STN DBS (Tröster et al., 1997; Ardouin et al., 1999; Gerschlager et al., 1999; Vingerhoets et al., 1999; Pillon et al., 2000; Daniele et al., 2003), even after 3 years (Funkiewiez et al., 2004), or only slightly declines even after 5 years (Contarino et al., 2007), other studies have found declines in working memory and response inhibition performance (Hershey et al., 2004), impaired frontal executive function, particularly in older patients, reduced verbal fluency, impaired naming speed, reduced selective attention, and delayed verbal recall (Saint-Cyr et al., 2000; Woods et al., 2006). Although the presence of dementia is considered a contraindication to DBS, cognitive function apparently improved in a patient with PD dementia with bilateral stimulation of the nucleus basalis of Meynert and STN (Freund et al., 2009). Further studies, however, are necessary before this approach can be recommended for patients with PD and cognitive deficit.

Several studies have reported changes in mood and behavior after DBS. Emotional lability and psychiatric complications, found in about 9% of patients after STN DBS, have been found in one prospective study with a control group consisting of PD patients without DBS (Smeding et al., 2006). In another study, bilateral STN DBS was associated with moderate improvement in tasks of prefrontal function and obsessive-compulsive behavior (Mallet et al., 2002, 2008) but moderate deterioration in verbal memory and prefrontal and visuospatial function (Alegret et al., 2001). While some patients have experienced euphoria, mania, infectious laughter, and hilarity with STN DBS (Krack et al., 2001; Kulisevsky et al., 2002), others have experienced depression and other mood changes, aggressive behavior (posteromedial hypothalamus), pseudobulbar crying (Okun et al., 2004b), and other psychiatric problems (Bejjani et al., 2002; Berney et al., 2002; Mayberg and Lozano, 2002), and some have noted improvement in psychiatric symptoms, particularly obsessive-compulsive disorder (Mallet et al., 2002, 2008; Tye et al., 2009). STN DBS may occasionally lead to increased impulsivity, similar to pathological gambling seen in patients treated with dopamine agonists, which may possibly be due to impaired ability to self-modulate decision processes and cognitive control (Frank et al., 2007; Stamey and Jankovic, 2008; Hälbig et al., 2009). The neuropsychologic and neuropsychiatric issues of DBS have been of increasing interest and concern and have been the topic of several reviews (Parsons et al., 2006; Saint-Cyr and Albanese, 2006; Voon et al., 2006; Appleby et al., 2007; Heo et al., 2008; Witt et al., 2008; Le Jeune et al., 2009). A meta-analysis of 10-year experience with DBS has concluded that the prevalence of depression was 2–4%, mania 0.9–1.7%, emotional changes 0.1–0.2%, and suicidal ideation/suicide attempt 0.3–0.7% (completed suicide rate was 0.16–0.32% over 2.4 years, compared to 0.02% annual suicide rate in the United States) (Appleby et al., 2007). Other studies have drawn attention to apathy (Le Jeune et al., 2009) and suicidal behavior as a potential hazard of STN DBS (Soulas et al., 2008; Voon et al., 2008). A survey of 75 DBS surgical centers representing four continents, 55 of which responded, indicated that out of a total of 5311 patients, 24 (0.45%) completed and 48 (0.90%) attempted suicide (Voon et al., 2008). The mean interval after surgery for all events was 17.8 months (range 1–48 months for completed suicides, 0.25–100 months for attempts), and 75% of all events occurred within this interval. The suicide rate in the first postoperative year was approximately 10 times the expected rate, adjusted for age, gender, and country, and remained elevated at 4 years. The risk of suicide correlated with postoperative depression, single marital status, and impulse control disorder. Although the rate of suicide following epilepsy surgery is even higher (1%), according to the authors “the baseline suicide rate [among patients with epilepsy] is eight times higher than the general population, [while] baseline Parkinson’s disease suicide rates range from the same to as much as ten times lower than the general population.” Transient mania also has been reported as an adverse effect of DBS, particularly involving the medial portion of the STN (Raucher-Chéné et al., 2008). STN DBS has been found to also improve autonomic dysfunction, including orthostatic hypotension, associated with PD (Stemper et al., 2006). Psychogenic tremor following DBS surgery for ET has been reported in one case (McKeon et al., 2008).

Hypophonia and dysarthria were reported to be the most frequent long-term effects of STN DBS (Romito et al., 2002). Dysarthria, observed as a side effect more with left than right STN DBS (Santens et al., 2003), may develop despite improved oral control of jaw, lips, and tongue movements (Pinto et al., 2003). While some individual components of speech may improve with STN DBS, most studies have shown that speech either fails to improve or deteriorates with STN DBS (Klostermann et al., 2008).

Several studies have now demonstrated a significant weight gain in patients with STN DBS (Montaurier et al., 2007). In one study, there was a 3.4 ± 0.6 kg body weight increase, associated with a significant decrease in energy expenditures, −7.3 ± 2.2% in men and −13.1 ± 1.7% in women (P < 0.01), associated with a gain primarily in fat-free mass in men and fat in women. In another study 32.1% of patients showed about 15% increase in body weight and a mean body mass index increase of 24.7 kg/m² in 1 year (Barichella et al., 2003).

While over half of patients initially considered to be DBS failures eventually had good results, 34% had persistently poor outcomes despite optimal management (Okun et al., 2005). In one study of 100 patients who were implanted with a total of 191 STN DBS devices, there were 7 (3.7%) device infections, 1 cerebral infarct, 1 intracerebral hematoma, 1 subdural hematoma, 2 (1%) skin erosions, 3 (1.6%) periprocedural seizures, and 6 (3.1%) brain electrode revisions (Goodman et al., 2006). There were 13 (6.8%) patients with postoperative confusion and 16 (8.4%) had battery failures, but there were no surgical deaths or permanent new neurologic deficits. The frequency of intraoperative complications associated with implantation of DBS was found to be 4.2% (11/262) and during a 3-year follow-up the frequency of hardware-related complications was 13.9% (25/180) (Voges et al., 2006). A total of 319 patients underwent DBS implantation at Baylor College of Medicine, Houston, Texas over a 10-year period, 182 of whom suffered from medically refractory PD; the other patients had essential tremor (113), dystonia (18), and other hyperkinetic movement disorders (6) (Kenney et al., 2007b) (Table 7.6). Intraoperative adverse effects were rare: vasovagal response in 8 patients (2.5%), syncope in 4 (1.3%), severe cough in 3 (0.9%), transient ischemic attack in 1 (0.3%), arrhythmia in 1 (0.3%), and confusion in 1 (0.3%). Perioperative adverse effects included headache in 48 patients (15.0%), confusion in 16 (5.0%), and hallucinations in 9 (2.8%). The most serious intraoperative/perioperative adverse effects occurred in 4 (1.3%) patients with an isolated seizure, 2 (0.6%) patients with intracerebral hemorrhage, 2 (0.6%) patients with intraventricular hemorrhage, and 1 (0.3%) patient with a large subdural hematoma (Kenney et al., 2007b). Long-term complications of DBS surgery included dysarthria (4.0%), worsening gait (3.7%), cognitive decline (4.0%), and infection (4.4%). Hyperhidrosis has been reported in a single case as a complication of thalamic stimulation for essential tremor (Diamond et al., 2007a). Revisions were completed in 25 (7.8%) patients for several reasons: loss of effect, lack of efficacy, infection, lead fracture, and lead migration. In another study of 60 patients who underwent 96 DBS-related procedures, followed over a period of 43.7 months (range 6–78 months), 18 (30%) developed 28 adverse events, requiring 28 electrodes to be replaced (Paluzzi et al., 2006b). The rate of adverse events per electrode-year was 8%. Hardware-related complications included 12 lead fractures and 10 lead migrations. In one study 25.6% of all 215 patients with DBS presented to the emergency department (ED) at least once for various reasons including neurologic (54.6%) such as headache (22.1%), change in mental status (15.1%), and syncope (9.3%), followed by infections/hardware issues (27.9%), orthopedic/focal problems (10.5%), and medical issues (7%) (Resnick et al., 2010). Several other studies have addressed DBS-related complications (Morishita et al., 2010).

Table 7.6 Safety data related to deep brain stimulation at Baylor College of Medicine and The Methodist Hospital

From Kenney C, Simpson R, Hunter C, et al. Short-term and long-term safety of deep brain stimulation in the treatment of movement disorders. J Neurosurg 2007;106:621–625.

Knowledge and skills needed to prevent and troubleshoot problems, such as a non-ideal initial DBS candidate, inadequate multidisciplinary team care, failure of perceived expectations, DBS procedural complication, hardware complication, suboptimal lead placement, programming, access to care, disease progression, and tolerance/habituation, are critical to the optimal response to DBS and to prevent potential “DBS failures” (Okun et al., 2008). Postoperative issues including outcomes and complications have been summarized in recent reviews (Deuschl et al., 2006a; Kleiner-Fisman et al., 2006; Voges et al., 2007; Hariz et al., 2008b; Videnovic and Metman, 2008).

Chronic stimulation appears to be well tolerated, and the risk of local gliosis is minimal (Caparros-Lefebvre et al., 1994; Haberler et al., 2000; Henderson et al., 2002), although more extensive damage has been rarely reported (Henderson et al., 2001). Electron microscopy examination of tissue adherent to the explanted electrode revealed foreign body multinucleate giant cell-type (some larger than 100 µm) reaction, possibly representing a response to the polyurethane component of the DBS electrode (Moss et al., 2004). Very few studies have addressed the rate and type of complications associated with DBS. During a mean follow-up period of 33 months and the cumulative follow-up of 217 patient-years of 84 consecutive cases, Oh and colleagues (2002) found that 20 patients (25.3%) had 26 hardware-related complications involving 23 (18.5%) of the electrodes. These comprised 4 lead fractures, 4 lead migrations, 3 short or open circuits, 12 erosions and/or infections, 2 foreign body reactions, and 1 cerebrospinal fluid leak. In addition, the hardware-related complication rate per electrode-year was 8.4%.

Several studies have addressed the safety concerns related to performance of MRI scans in patients with implanted DBS devices. Standard and functional MRI scans have been performed in many patients with implanted DBS without significant complications (Jech et al., 2001; Tagliati et al., 2009) (Table 7.7). In a survey of medical directors of National Parkinson Foundation Centers of Excellence, data on MRI safety was available on 3304 PD patients with one or more DBS leads; DBS patients had MRI of other body regions. No complications were reported except in one case in which MRI of the brain was associated with an IPG failure with no neurologic sequelae (Tagliati et al., 2009). Diathermy treatment, which involved pulse-modulated radiofrequency to the maxilla, in a 70-year-old patient with PD implanted with ITREL model 7424 in the STN, resulted in permanent diencephalic and brainstem lesions and a vegetative state (Nutt et al., 2001a). This tragic complication probably resulted from induction of radiofrequency current and heating of the electrodes, leading to the edema surrounding the DBS electrode. Although no serious complications have been reported in women with implanted DBS during pregnancy or delivery (Paluzzi et al., 2006a), the use of electrocautery during delivery by cesarean section should be used cautiously.

Table 7.7 Results of a safety survey on MRI in patients with DBS devices

From Tagliati M, Jankovic J, Pagan F, et al.; National Parkinson Foundation DBS Working Group. Safety of MRI in patients with implanted deep brain stimulation devices. Neuroimage 2009;47(Suppl 2):T53–T57.

To improve the safety and efficacy of DBS, it is important to continue advancing our understanding of the mechanisms underlying DBS and to build on and improve DBS technology (e.g., develop directional electrodes, prolonging battery life, allow remote and “smart” programming).

Dystonia

GPi DBS also has been reported to be effective in patients with primary generalized dystonia (Krauss et al., 1999; Kumar et al., 1999a; Coubes et al., 2000; Tronnier et al., 2000; Brin et al., 2001; Krack and Vercueil, 2001; Muta et al., 2001; Vercueil et al., 2001; Albright, 2003; Coubes et al., 2004; Diamond et al., 2006; Vidailhet et al., 2005; Jankovic, 2006; Kupsch et al., 2006; Jankovic, 2007; Vidailhet et al., 2007; Kenney and Jankovic, 2008; Isaias et al., 2008, 2009), segmental dystonia (Wohrle et al., 2003), cervical dystonia (Bereznai et al., 2002; Krauss et al., 2002), blepharospasm-oromandibular (cranial) dystonia (Capelle et al., 2003; Foote et al., 2005; Vagefi et al., 2008), tardive dyskinesia (Eltahawy et al., 2004a; Lenders et al., 2005), and tardive dystonia (Trottenberg et al., 2001b, 2005; Cohen et al., 2007; Sako et al., 2008; Gruber et al., 2009). Vercueil and colleagues (2001) described ten patients with bilateral GPi DBS; five had a major improvement, two had a moderated improvement and one had a minor improvement after 14 months. They concluded that GPi DBS is much more effective than thalamic DBS for the treatment of dystonia. In a 2-year follow-up of 31 patients with primary generalized dystonia, Coubes and colleagues (2004) noted a mean improvement in the clinical and functional Burke–Fahn–Marsden Dystonia Rating Scale (BFMDRS) of 79% and 65%, respectively. There was no difference in response between DYT1-positive and DYT1-negative patients, but the magnitude of improvement was greater in children than in adults. In a presentation during the seventh International Congress of Parkinson’s Disease and Movement Disorders in Miami in November 2002, Coubes provided data on 68 patients with generalized dystonia, 78% with primary dystonia, 19 of whom were DYT1-positive, and 29% with secondary dystonia, treated with GPi DBS. The MRI-guided, single-tract, bilateral implantation showed an overall 80% improvement, and this was most robust in primary, DYT1 dystonias and in the pantothenate kinase-associated neurodegeneration cases. Primary dystonia clearly responds better than secondary dystonia to either pallidotomy or GPi DBS, although patients with pantothenate kinase-associated neurodegeneration also experience marked improvement in their dystonia (Coubes et al., 2000; Eltahawy et al., 2004b; Castelnau et al., 2005). Essentially all patients achieved steady state in 6 weeks. The improvement was particularly noticeable in the rapid (“ballistic”) dystonic movements, pain, and bradykinesia, but there was no effect on the slow dystonia or associated bradykinesia. Except for implantable pulse generator infection in three patients and lead fracture and other lead problems in two patients, there were no other complications. The stimulating parameters were as follows: pulse rate was 450 µs, frequency was 130 Hz, and amplitude was 0.8–1.6 V. In a prospective, multicenter study of 22 patients with generalized dystonia, 7 of whom had DYT1 mutation, the BFMDRS score improved after bilateral GPi DBS from a mean of 46.3 ± 21.3 to 21.0 ± 14.1 at 12 months (P < 0.001) (Vidailhet et al., 2005). A “blinded” review of the videos at 3 months showed improvement with stimulation from a mean of 34.63 ± 12.3 to 24.6 ± 17.7. The improvement in mean dystonia motor scores was 51%, and one-third of the patients improved more than 75% compared to preoperative scores. In addition, there was a significant improvement in health-related quality of life as measured by the SF-36, but there was no change in cognition or mood. Although the sample was rather small, the authors were not able to find any predictors of response such as DYT1 gene status, anatomic distribution of the dystonia, or location of the electrodes. Patients with a phasic form of dystonia improved more than those with tonic contractions and posturing. The maximum benefit was not achieved in some patients until 3–6 months after surgery. In a 3-year follow-up, motor improvement observed at 1 year (51%) was maintained at 3 years (58%) and the authors concluded that “bilateral pallidal stimulation provides sustained motor benefit after 3 years” (Vidailhet et al., 2007). In a study involving French centers, bilateral DBS of ventral GPi was associated with a 42% reduction in the BFMDRS score; whereas DBS of dorsal GPi resulted in less predictable effects (Houeto et al., 2007). In a retrospective review of 40 patients with primary generalized dystonia who had been treated by bilateral GPi DBS followed for up to 8 years, the most important predictors of poor response were high preoperative BFMDRS score, older age at surgery, and small GPi volume, but no significant correlation was found between the electrical parameters used and the mean motor scores (Vasques et al., 2009).

Neurophysiologic studies at the time of the implantation of stimulating electrodes or during chronic stimulation have provided insights into the pathophysiology of dystonia and how DBS alleviates the involuntary muscle contractions. Based on a study of 15 patients with primary dystonia treated with bilateral pallidal stimulation, modulation of oscillatory local field potentials (LFPs) were recorded from pallidal electrodes and were correlated with surface electromyography of the affected muscles (Liu et al., 2008). Dystonic movements were associated with increased theta, alpha and low beta activity and the strength of the contraction correlated with an increase in frequency range of 3–20 Hz; the increase preceded the spasms by about 320 ms. There was a significant decrease in LFP synchronization at 8–20 Hz during sensory modulation, but voluntary movement increased gamma band activity (30–90 Hz). It has been suggested that DBS alleviates the hypertonic activity by desynchronizing these excessive synchronized discharges. In six patients treated for their dystonia with bilateral GPi DBS, the contralateral prefrontal overactivity was reduced (Detante et al., 2004). Kupsch et al. (2006) compared GPi DBS with sham stimulation in a randomized, controlled clinical trial of 40 patients with primary segmental or generalized dystonia. The primary endpoint was the change from baseline to 3 months in the severity of symptoms, according to the movement subscore on the BFMDRS. Two investigators who were unaware of treatment status assessed the severity of dystonia by reviewing videotaped sessions. Three months after randomization, the change from baseline in the mean (±SD) movement score was significantly greater in the neurostimulation group (−15.8 ± 14.1 points) than in the sham-stimulation group (−1.4 ± 3.8 points, P < 0.001). Similar results were obtained in other centers (Krause et al., 2004).

Targeting the posteroventral GPi seems to provide the most robust benefits in patients with dystonia (Tisch et al., 2007). It is of interest that an ablative lesion or high-frequency stimulation of the GPi can both produce and improve dystonia, suggesting that it is the pattern of discharge in the basal ganglia rather than the actual location or frequency of discharge that is pathophysiologically relevant to dystonia (Münchau et al., 2000). In one patient, a 49-year-old woman with severe generalized dystonia, bilateral GPi DBS produced an immediate improvement in dystonia, which was associated with a reduction in PET activation in certain cortical motor areas that are usually overactive in dystonia (Kumar et al., 1999a). Bilateral GPi DBS has been found effective not only in distal or generalized dystonia, but also in patients with cervical dystonia (Krauss et al., 1999; Kulisevsky et al., 2000; Parkin et al., 2001). In a report of two patients with cervical dystonia, bilateral GPi DBS produced more improvement in pain than in motor symptoms (Kulisevsky et al., 2000). Parkin and colleagues (2001) reported progressive improvement in pain and posture in three patients with cervical dystonia. A prospective, single-blind, multicenter study assessing the efficacy and safety of bilateral GPi DBS in 10 patients with severe, chronic, medication-resistant cervical dystonia found that the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS) severity score improved from a mean (SD) of 14.7 (4.2) before surgery to 8.4 (4.4) at 12 months postoperatively (P = 0.003) (Kiss et al., 2007). The disability and pain scores also improved, as did general health and physical functioning and depression scores.

GPi DBS has been also reported to improve cranial-cervical dystonia in selected cases (Ostrem et al., 2007; Blomstedt et al., 2008; Pretto et al., 2008; Markaki et al., 2010). While bilateral GPi DBS may improve symptoms of dystonia, motor function in non-dystonic body parts may worsen (Ostrem et al., 2007). Muta and colleagues (2001) described a 61-year-old woman with cranial dystonia manifested chiefly as blepharospasm, facial grimacing, cervical dystonia, and spasmodic dysphonia who had previously failed to improve with bilateral thalamotomy but who had marked improvement in all aspects of her dystonia with bilateral GPi DBS, including complete resolution of blepharospasm and oromandibular dystonia. In a blinded review of videos of 13 patients with segmental dystonia caused by various etiologies and different distributions (4 with cranial dystonia), GPi DBS was associated with global subjective gains and notable objective improvement in 11 of 13, but the response was quite variable and unpredictable (Pretto et al., 2008). The pattern of recurrence of segmental dystonia after discontinuation of GPi DBS was studied in 8 patients (Grips et al., 2007). Phasic dystonia appeared within a few minutes but the tonic form of dystonia recurred with a more variable delay. Some patients obtain optimal improvement with lower-than-usual stimulation frequency (80–100 Hz) (Alterman et al., 2007). GPi DBS is gaining acceptance as the surgical treatment of choice not only in adult patients with dystonia but also in children (Alterman et al., 2007).

Tourette syndrome

There has been increasing interest in DBS as a treatment of medically intractable or malignant Tourette syndrome (TS) (Cheung et al., 2007). Because of prior success with thalamic ablation in the treatment of severe TS reported in 1970 by Hassler and Dieckman, recently translated from French (Rickards et al., 2008), most investigators use the original report as a rationale for selecting the thalamus as the target in ablative and DBS treatment of TS. In 1999, thalamic DBS was introduced for intractable TS, but since then, multiple targets have been used with relatively comparable results (Ackermans et al., 2008). Globus pallidus has been increasingly used as the target in patients with disabling tics because of the long-term experience with this target in the treatment of other hyperkinetic movement disorders, such as dystonia and levodopa-induced dyskinesia (Diederich et al., 2005). In a 16-year-old boy with disabling, medically intractable TS, bilateral GPi DBS resulted in 63% improvement in Yale Global Tic Severity Scale, 85% improvement in Tic Symptom Self Report, and 51% improvement in SF-36, a quality of life measure (Shahed et al., 2007). Furthermore, the patient was able to return to school. Several other reports confirmed that GPi is an effective target for the treatment of TS (Dehning et al., 2008). Although these observations must be confirmed by a controlled trial before DBS can be recommended even to severely affected patients, it suggests that stimulation of certain targets involved in the limbic striato-pallidal-thalamo-cortical system may be beneficial in the treatment of various aspects of TS. In TS, perhaps even more importantly than in other movement disorders, appropriate screening, and accurate and comprehensive assessment of not only tics but also behavioral comorbidities is absolutely critical in patient selection for DBS (Mink et al., 2006). Surgery for treatment of TS is discussed in more detail in Chapter 16 (Porta et al., 2009; Ackermans et al., 2011).

Other Indications for DBS

The expanding use of DBS in the treatment of various disorders could provide insights into the pathophysiology of other disorders (Benabid et al., 2000). In addition to the hyperkinetic movement disorders already discussed, DBS may be useful in patients with severe chorea associated with Huntington disease. Bilateral GPi was found to control chorea and markedly improved the quality of life in a 60-year-old man with a 10-year history of Huntington disease (Biolsi et al., 2008). Bilteral GPi DBS has been also reported to produce about 24.4% improvement in the Burke–Fahn–Marsden scale in adult patients with dystonia–choreoathetosis associated with cerebral palsy (Vidailhet et al., 2009). It has been also found to be helpful in patients with neurodegeneration with brain iron accumulation (NBIA), although not as much as in patients with primary dystonia (Timmermann et al., 2010).

In addition to VIM as a target for patients with disabling tremor, STN DBS has been also found to be useful in the treatment of severe proximal tremor (Kitagawa et al., 2000) as well as rest and postural tremor in PD (Sturman et al., 2004).

The observation that stimulation of the substantia nigra precipitates acute depression (Bejjani et al., 1999) or mania (Ulla et al., 2006) that resolves immediately when the DBS is turned off suggests that the nigrothalamic pathway may play an important role in bipolar disorder. Besides left substantia nigra stimulation (Bejjani et al., 1999), stimulation superior and lateral to the right STN also produced mood changes (dysphoria) (Stefurak et al., 2003). Stimulation of white matter adjacent to the subgenual cingulated region (Brodmann area 25), which is metabolically overactive in treatment-resistant depression, has been found to be effective in treating depression in four of six patients (Mayberg et al., 2005).

Other emerging applications of DBS include treatment of obsessive-compulsive disorder targeting chiefly the ventral anterior limb of the internal capsule and adjacent ventral striatum or STN (Gabriels et al., 2003; Nuttin et al., 2003; Mallet et al., 2008; Greenberg et al., 2010), as well as various pain disorders, including migraines and cluster headaches (Leone et al., 2003) (Table 7.8).

Table 7.8 Current and potential indications for deep brain stimulation